Category: Uncategorized

How A.I. just changed the Internet

Posted on by Steve Sammartino

Welcome to 2023 – This year I’ve promised myself to post here every Friday – starting today of course.

The impact of ChatGPT has been well documented in the media. If you haven’t played around with it yet – then I strongly suggest you give it a whirl.

There have been many articles about which industries and jobs ChatGPT might disrupt. Here’s a much simpler way to look at it from the Sammatron: If your job or industry uses ‘words’ – it will be impacted.

It’s that simple. It will impact everything. Like the internet that came before it, ChatGPT brings inordinate opportunities for those who choose to embrace the new reality.

The internet is currently:

A Giant Filing Cabinet

Everything we see on the internet is made by someone, somewhere. Words, pictures and video, all of it. Our experience on the internet is basically asking services to find and serve up what we’ve subscribed to or what we’ve searched for in this giant filing cabinet of human creation. Even when it is live, it’s still us peering in the window of what someone else is making or doing. This is why search and social are so powerful. That’s about to change.

We are about to go from Search Engines to Creation Engines.

(Read the above sentence again, it’s that important)

Let me explain. Starting now – we will shift from asking the internet to find what we want to asking it to create what we want. We’ll be expecting it to answer exact questions to queries, summarise things, write for us, create images, pictures, stories, animations, rap songs and even write software code for us by literally asking it to create an app that does XYZ. We are about to see another level of technology democratisation beyond what anyone can imagine.

ProTip: Get me to come into your company and deliver my mind blowing new Keynote Speech on How AI is about to change everything.

The art of the internet will rely on our ability to work with Artificial Intelligence. We’ll need to cajole the AIs to create what we are after. We’ll need to become great at prompting and teaching AIs how to build what we desire, because when used correctly, AI should be a mind expanding tool. An addendum to our biology, not a replacement. We’ll need to learn how to literally ‘train the AIs’ like we might a train a dog. It will be about the symbiosis of using tools. From a business perspective, our attention should be focused on industries that can cut costs through large language models and AI-generated imagery. The business of words and images will have production costs cut dramatically.

Get Daily Inspiration on my Tik Tok – 4 million views & counting! – Follow me here

Our imaginations on what to ‘make’ will become increasingly important. Here’s a prompt we might give very soon: “Create a seven minute animated movie, in manga style. Let it star a seven year old girl named Mary (base her animation on this picture) and her magic pony, who saves the world from a climate crisis. Give me versions in English, Spanish and Mandarin.”

As  I’ve previously written, the ability to determine what was made by a human or a machine will be increasingly difficult to differentiate. Likewise, an entirely new species of startups and business opportunities will emerge. It’s early, so it’s time to get started.

– – –

Keep Thinking,

Steve.

MetaVerse – Use cases

Posted on by Steve Sammartino

The MetaVerse will definitely be a thing – a big thing in fact, just not a social thing. Here’s why:

The Metaverse has all the inconvenience of catching up in real life without any of the benefits.

Below I’ve listed ways industries can, will and already are employing Metaverse applications. So here are my best 3 use cases of the Metaverse. Hopefully it will help you understand, if and when, your company or industry should embrace the tech.

(1) Gaming & Entertainment: 

Video gaming is no small industry with an expected revenue of over $200 billion in 2022. It’s already larger than Hollywood and filled with significant competitors including Sony, Microsoft, Nintendo, Tencent, Electronic Arts and Epic Games, all of which used virtual and mixed reality and have large and loyal gaming populations. While many of these games have a multiplayer component, they aren’t driven by social interaction, it is more a side benefit. This sector will continue to morph into the Metaverse. We can expect movie studios to also enter this space, especially as video streaming eats into distribution networks – and they try and differentiate the cinema experience by making it more immersive.

(2) Training & Education: 

While pilots have long used simulated reality for flying training, VR is now starting to be used for surgeons and other professions which require physical dexterity and risk reduction during the training process. We could even use it to assist learning to play and instrument, or learning a language. Eventually trade schools will use Metaverse style applications.

BONUS: Now open – Emerging Trends session for 2023 – Click here. (special offer to subscribers)

(3) Computer Aided Design:

 I was recently at Rio Tinto and took a virtual tour of a mine which was to be constructed. It felt incredibly real, so much so that I continued to duck my head while walking around. We can expect the metaverse to be used as a pre-production tool before anything of significant size is built. Think factories, warehouses, hotels and even our homes. Again, this space is well occupied by incumbents like AutoDesk and not particularly social in nature.

This will become an incredibly interesting and large business, just not very social in nature. But the short term use cases are actually quite clear and already filled with highly capable incumbents.

It seems Zuckerberg’s bet in this as a social application is way off. And if the $700 billion market valuation loss and 11,000 staff retrenchments this week are any indication – Meta (Facebook) as a business, is in for a bumpy ride.

—

Keep Thinking,

Steve.

Metaverse – Hype of Reality?

Posted on by Steve Sammartino

Why Now?

When a top 10 company by valuation like Facebook, changes its name and focus to the MetaVerse – it is going to generate a lot of attention. A google search of the term now has 141 million results, a more than 100 fold increase on a year ago. The reason that Facebook became Meta, and started its focus on the concept, is that Facebook after being one of the fastest growing companies in history, is now in decline. In terms of both revenue and profit.

Since its launch in 2004, Facebook was the clear winner of social networking. It has 3.6 billion users across its platforms which include Facebook, Instagram and WhatsApp. But now, it is starting to wane as TikTok encroaches and people turn away from its apps. Meta wants to own what it thinks will be the future of social interaction. Meta’s revenue is almost entirely dependent on advertising (97%). Their success in advertising is in no small part due to the information they can gather on their users. Meta knows what you like, do, believe, we’re you’ve been, where you’re going next, who you friends are, what your life history is and of course, and what you spend on. And it’s not just through their sites, rather via the likes and login ecosystem they’ve built on the web which creates such a rich data flow.

BONUS: How observant are you? Have you noticed this?

There is no organisation in the history of the planet who nows more about individuals. They know more about you than your partner or government does. But this information gathering has become far more difficult as the owners of the hardware people access Meta apps on (namely Apple and Google) are starting to restrict Meta’s ability to track users. Likewise, governments the world over are quickly regulating against these privacy incursions, which are the fuel that powers the organisation.

This is a major problem for Meta, as don’t own any significant hardware. The Meta portal device has been discontinued and I don’t know anyone with a pair of their RayBan spy glasses. Because the Metaverse requires the use of goggles like the Meta Quest 2, it could solve this hardware problem and allow Meta to set its own privacy terms. To date they’ve made a massive investment including a $2 billion acquisition of the Oculus company and more than $70 billion in developing Metaverse hardware and applications. Zuckerberg really needs this to work.

BONUS: Start Smart in 2023?: Get the Sammatron in to speak to your team – Special offer for new year bookings to blog readers.

Hype or Reality?

So far, the share market hates it. The Meta Share price has declined more than 60% in the past year, losing investors more than $700 billion since its peak. That’s not a typo. In short, Mark Zuckerberg needs the Metaverse more than we do. But it is hard to see 3.6 billion people going from a ‘free service’ to having to invest $600 plus into a single purpose device (MetaQuest VR goggles) to log into Facebook Horizons.

In my view this is why the Metaverse is getting such attention. The only question is whether the people of the internet themselves will give it as much attention as Zuckerberg wants them to. If the meagre house hold penetration of VR goggles being at 0.03% of Facebook users is any indication – the answer is a clear no. Or at least, not for a long time.

Next Week: MetaVerse – Best use cases.

—

Keep Thinking,

Steve.

The Metaverse Lowdown

Posted on by Steve Sammartino

It’s been almost a year since Facebook changed its corporate name to Meta Platforms. In that time, not unexpectedly, the ‘Metaverse’ has garnered much attention. I’m frequently asked by clients what I think about the Metaverse and whethe they should get on board. So here it is, from the Sammatron: the low-down on the Metaverse, where it’s at and where it might go next. I’ve started with a simple explanation of the Metaverse:

Metaverse 101

Firstly, the Metaverse isn’t about to ‘arrive’ like the iPhone did. It’s already here. Second Life, Minecraft, Roblox and all virtual reality applications can be considered Metaverses. In order to understand what the Metaverse is, we need to think of as a continuation of the internet. It’s an iteration, but one that forks into a different direction.

The concept of a Metaverse is built on the idea that one day we will live ‘inside’ computer systems. We can already see this trajectory looking back at the evolution of the web. At first, we had text only green screens with a limited number of connected computer systems to share information. Eventually we invented message boards, then hyperlinks and web browsers with graphical user interfaces. From here the web became an image repository with gifs, still images and high resolution video with real time livestreams. But all of it has been screen-oriented, with a 2D audio visual focus. Just like the smart phone was a fork in direction based on a new device, future Metaverse experiences are expected to become dependent on immersion devices such as virtual reality goggles.

The Metaverse aims to provides a whole body experience. First with headsets and glasses, and eventually haptic suits and gloves, so it feels real to your entire body. Virtual reality goes some of the way toward this experience, providing a 3D style reality that creates a sense of immersion. But the plan for those pushing the Metaverse is that all five humans senses will be involved with the ability to touch, feel, hold, walk and even smell and taste inside artificial environments. These more sophisticated environments will allow for interactions not currently possible on the internet. It would allow the web to cross the chasm and make many things feel real, rather be limited to on-screen replicas. The promise is that basic websites will have a Metaverse version of them, like many www sites already have app versions of them. One example is an ecommerce site becoming a virtual store you can interact inside virtually. It’s also touted that we’ll work and socialise inside the Metaverse. Just like apps and the web, there won’t be only one, but a multitude of Metaverse locations we can enter. In the long run, we won’t even need goggles. Instead we can connect to virtual worlds via brain machine interfaces (think Matrix). If this happens (I think it will – in decades’ time) the difference between the real and virtual world will be very difficult to distinguish.

Next week: The Metaverse hype versus reality.

—

Keep Thinking,

Steve.

TV Reality Check

Posted on by Steve Sammartino

Who’s the biggest TV star in the world? Anyone older than 25 will probably nominate a TV star from the free-to-air (FTA) TV era. Anyone under the age of 25, that is, born after the modern internet burst onto the scene, will tell you it’s MrBeast – aka Jimmy Donaldson.

MrBeast recently passed a milestone 100 million subscribers on YouTube (now 104m) . In total, his posts have attracted more than 26 billion views and his videos are seen on average anywhere between 50 to 110 million times. To put that in context, the two most-watched shows in TV history (outside of global live events) are:

  • Mash – the final episode, 105 million viewers and;
  • Seinfeld – the final episode 76 million viewers

Clearly, it has been a long time since traditional linear TV set any viewership records.

Mr Beast Strategy

MrBeast is now more than an on-screen persona – it’s graduated to a sizeable business. According to Forbes magazine, Donaldson earned $US54 million in revenue last year and now runs a production company with over 50 employees. He has also given away more than $50 million to charity and claims to be YouTube’s biggest philanthropist. To celebrate his 100 millionth subscriber, Donaldson created a video with a hundred of his fans running a modern-day game show, competing to win an island. The production cost alone for this celebration was over $US500,000.

The story of Donaldson’s origin is as surprising as how big he’s become. Ten years ago, Donaldson set himself a clear goal to become famous by earning as many views on YouTube as possible. The ‘how’ of gaining views was less important. Naturally, Donaldson gravitated towards silly stunts and challenges to maximise reach. Some of his early videos included counting to 100,000 (a lengthy 40 hours) and tried spinning a fidget spinner for 24 hours straight. Although his videos have evolved in sophistication of production and concept, this ‘game show’ method has served him well. His key ingredients have become:

  • Challenges and stunts 
  • Giveaways with often very generous prizes
  • Celebrity appearances

Consistent with other popular game shows like The New Price is Right, audience participation is a priority. Donaldson creates incentives for viewers to be involved – to appear in videos and compete for exciting prizes, creating a brilliant viral loop growing both views and audience. Donaldson claims to have cracked the code for making viral posts and that once you’ve mastered the art, “you can make unlimited money”. There are many lessons our local media channels could learn from him about generating views, revenue and profit.

Bonus: Follow my TikTok here for biz & life rad advice (almost 1 million views just this month!)

TV Reality Check

The financial struggles of free to air TV channels are well-documented. Network 10 ended up in receivership and was eventually bought out by CBS. Neither Seven West Media nor Nine Entertainment have done much to solidify their long-term future and may struggle.

Basically, all media sells attention. In media circles, this is known as CPM which is ‘cost per thousand’ viewers. Enter, a reality check. FTA TV currently commands an average premium per set of eyeballs of 700 per cent. No, that’s not a typo.

Nestled in the TV that fits in the palm of your hand, social media platforms including Instagram, Facebook and YouTube charge around $A7 for every 1,000 views. In comparison, FTA TV charges $A50 per 1,000 views a mid-rating show, while ratings winners like The Block costs as much as $175 per 1,000 viewers for a 30 second commercial.

The irony is that while most people are peering into their tiny phone screens, that is when these ‘premium-priced’ commercials run on the big TV mounted on the wall. TV commercials are not nearly as targeted as digital advertising is, neither psycho-graphically nor geographically.

Additionally, FTA TV commercials don’t have click-through potential, and disappear into the ether after they run. In contrast, the online variant, while it can be skipped, can continue to live on in the same social channel as a standalone video, receiving organic views and feed a content play for the brands pushing them.

These FTA numbers for reaching people simply can’t and won’t be sustained. If I were to point out a reason why the market still tolerates such an imbalance in CPM on FTV versus digital channels, I’d narrow it down to the senior executives buying the ad spaces. People buying the advertising space for large brands mostly grew up before the internet and bear a legacy mindset that TV is in some way superior.

That said, it won’t be long before a new cohort of CEOs and CMOs arrive and start challenging why they are paying such an exorbitant premium to reach the same people. When that happens, we ought expect FTA TV’s revenue to decline by 80 per cent.

What Free to Air TV Should Do

Like all disrupted businesses, traditional TV is filled with growth opportunities if they’d only embrace them. The first of these should be realising that the world is no longer segmented in 30-minute slots.

The simplicity of online videos running for as long as a few seconds to a few hours is telling. It seems as though the TV stations would rather things be neat, instead of relevant.

They need to lower the barriers to entry on their digital catchup TV. Tuning into catchup TV is not nearly as easy as it should be. It requires registrations and most shows are removed after a few weeks. In what should be a competitive play against YouTube, the on-demand digital platform is treated as a departures lounge for the TV you missed.

In fact, all channels would do well to leave their entire corporate back catalogue of TV shows online forever, share the revenue with existing content owners and even allow it to be mashed up by new content creators. Then they could create new layered versions of what has already been made. This is already happening, but just TikTok and YouTube. It’s not even a secret…. they only need to copy what is happening!

Search for your favourite TV show from the 1980s and you’ll see it published on YouTube with Google collecting the revenue for advertising running before and beside it. Simply embracing long tail content, asynchronous viewing and allowing the audience to upload and re-interpret their content could change Free to Air’s business model dramatically. But it seems they love the business model they had yesterday more than the reality of today.

– – –

Keep Thinking,

Steve.

The Attention Scoreboard

Posted on by Steve Sammartino

Scoreboards change everything. Once we have a metric in place, whatever improves that metric becomes the focus of what we do. We do it with sport, we do it with money, and now we do it with content. A metric you all know is how many views on average your social posts attract. We all do.

Attention wins

Once upon a time, great content was the focus. Content is king, we’d all chant ad nauseam. The inference was that the best stuff would naturally bubble up to the top. However, humans are weird and emotional creatures. We have certain neuroses we can’t help but serve. As a social species, we are infused with the need to be part of a group to survive. When others are engrossed in something new, of course our curiosity is piqued. In this era, we can see exactly how many people are paying attention to something. This scoreboard sits under every photo and video. It almost doesn’t matter what it is, so long as people are watching. Algorithms perpetuate that further. Something that is popular becomes more popular, because it is popular.

This presents a real challenge for people who want to produce content that is important, but perhaps not quite as entertaining. While a business may insist their primary goal is quality for their content, it’s hard not to be swayed even more by the number of views it receives. When Tommy and I launched our TV show, The Rebound, every sponsor told us their focus was delivering great insights and hacks to the audience, yet what they couldn’t resist asking first in every review would be, “How many viewers did each episode get”?

Know your metric

All this comes back to your ‘why?’ Why are you creating content? The truth is it isn’t that hard to go viral. I have a handful of YouTube videos that have clocked up well over a million views. We just need to appeal to the market of emotions and be as extreme as possible. If attention is the key metric, then we just follow where the attention economy takes us. Get on a bandwagon, roll the dice enough times and eventually you’ll score. But for what purpose? If entertainment and attention are your key metrics, then that’s exactly where your focus should be.

One the other hand, if you’re not aiming for the mass market, then deeper consideration must be taken.

I’m after an audience who want knowledge and seek consulting services and keynote speeches from me. We need to decide what we are chasing before we start creating. We need to resist the temptation to chase numbers, otherwise the attention scoreboard redefines what we do and who we are, with unintended consequences,.

—

Keep Thinking,

Steve.

From Big Tech to ‘Big Voice’

Posted on by Steve Sammartino

Language is humanity’s killer app. We are the only species that can communicate complex ideas and also pass them through generations. It’s often said that our ability to communicate in a such a manner has put us atop of the food chain. Big technology firms have become so powerful because they’ve put a metaphorical fence around all that we say, and therefore, control all that we know. Big tech as we’ve come to know them are basically, in the language business. 

While it’s clear that our voices all all unique. What is less known is what can be gleaned from them

The Verbal Reveal

Despite our hearing abilities being inferior to many other animals, we humans can deduce a surprising amount of information from listening to each other speak. We can make pretty good guesses at someone’s gender, age and perhaps even levels of education.

However, machines can infer far more than a human ever could from a voice. In the developing branch of Artificial Intelligence which focuses on speech, much has been achieved, even in the early days of understanding sentences and how to respond. Known as Natural Language Processing (NLP), this field of A.I. can discern with significant accuracy someone’s age, gender, ethnicity, education levels, socioeconomic status, and even uncover health conditions.

What we tend to forget is that the voice recognition system is doing more than just listening to words. It is also cross-referencing what has been said with other information that can be extracted from the device (smartphone / internet browser) delivering the voice data.

In essence, NLP does what humans do. When we meet with someone, we look at our surrounds to provide clues on how to navigate the conversation. We look at what people are wearing and the venue we are in – everything visual and contextual that might support the verbal interaction. Conversation that comes with context nearly always has better outcomes.

This is part of the reason telephone customer service lines are so often a poor end-user experience. When an NLP engine interacts with enough voices for long enough, it can create new forms of pattern recognition. Sourced from vast data sets incorporating 65 per cent of the world’s population, NLP can match data points to voice prints far beyond what any human ever could. What NLP may uncover is almost limitless.

The Mirror of the Brain

Using our voices is not as simple as it seems. When we speak, a complex process is launched, activating both physical and mental faculties. Involving the lungs, voice box, throat, nose, mouth, lips, sinus and jaw shape. Using your voice activates more than 100 muscles every time it is used. Additionally, your voice is the reflector of the brain – what it knows, believes and how it responds to audio stimuli. As the MIT Media Lab voice researcher Rébecca Kleinberger says: “The voice is very much the brain.”

Voice to Face

A single voice can now reveal unfathomable volumes of information to an A.I. engine. Researchers have even been able to generate images of faces based on information ascertained from individuals’ voice data. I know – this blew my mind too.

A 2019 Cornell University research study was able to reconstruct facial images of people using short audio recordings of their speech. The facial image reconstruction was produced through training a deep neural network, utilising millions of YouTube videos of people conversing naturally and without a script.

The network training methodology looked for correlations and co-occurrences of faces and voices. It matched the probability of voice patterns with pixel patterns to guess what an unknown person’s face might look like. The results below are quite astounding, given how nascent this technology is.

BONUS – Follow me on TikTok for daily motivational sound bites of goodness. No dancing from me – I promise.

Everlasting Conversations

At Amazon’s recent re:MARS conference for Machine Learning, Automation, Robotics and Space, a new Alexa feature was thrillingly unveiled. Amazon’s AI assistant can now impersonate voices of users’ dead relatives. The demonstration featured a child asking his deceased grandmother to read out a bedtime story, whereupon her voice obligingly pours out of a nearby speaker. As you can imagine, this feature was met with both admiration and outrage. Many called it plain creepy.

While we might think that it would take a deep longitudinal data set for Alexa to learn how to mimic a real human voice, Amazon claims that its AI system can learn to imitate a voice from less than a minute of recorded audio. Given how prevalent recordings of people are now on both video and audio, it wouldn’t be difficult to create a voice clone of a loved one. Or pretty much anyone who has ever been on the internet.

While Amazon hasn’t given an indication whether this feature will be rolled out, the technology will surely leak across the web, as it always does. The website Fakeyou.com is a veritable “Choose Your Own Adventure” of actors, celebrities, singers, cartoon characters and public figures whose voices can be manipulated to say whatever you desire. It’s possible to upload your own voice and even clone that.

The opportunities bubbling out of this are intriguing. When we have to make that dreaded phone call to the bank or tax department, we may be able to choose from Scarlett Johansen or Ryan Gosling as our customer services operator to make solving the issue a little more pleasant. If we combine already existing music AI systems with voice cloning we may be able to create new Beatles tunes – John Lennon may sing again. Animated voice actors for long-running animations like The Simpsons may well be delivered a pink slip. Radio hosts from years gone by could join the podcast bandwagon to cater to more senior audiences.

The Voice Cloning Industry

Voice is the latest in a long line of emergent cloning marketplaces. This creates new value, revenue sources, investment and of course, potential legal conflict. The voice and speech recognition industry (I’m calling it ‘Big Voice’ — you heard it here first) is estimated to exceed $US20 billion by 2026.

If we thought it was already difficult to distinguish between what is real and what is fake on the internet, it’s only going to get a lot more challenging.

—

Keep Thinking,

Steve.