Your June issue of Small Cap Investigator

Artificial intelligence: creation or destruction?

Last weekend, I got an email from Google’s AI Test Kitchen.

This is a branch of Google’s public access research and development into artificial intelligence (AI). Anyone can register for AI Test Kitchen – just search for it and register your interest.

You go onto a waitlist, but it will let you know when it’s your turn – you’ll then gain access to things that Google wants to test.

The latest thing I’ve been sent from the AI Test Kitchen is its “MusicLM Demo v0.01”.

It’s a music generator where you can create music simply by prompting the AI. For example, here is one prompt I tried out when I was granted access:

2000s style heavy rock, double kick, fast paced, melodic heavy chords with a notable bass riff I can work out to.

Two seconds later it made this:

The next prompt I tried was this:

Late 1990s house, meet her at the love parade style, with a horns section melody, and deep looping base I can dance to at 128bpm.

Two seconds later, it made this:

The music might not necessarily be to your taste, but it’s original music – and I created it. Or did I? I guess I prompted the creation, but who really made this music?

And perhaps the bigger question should be, who owns this music? Me? The AI? Or Google?

The reason I bring this up is for two reasons. Two competing reasons that signify everything that’s going on in the world today relating to AI. Two reasons that you need to understand to fully comprehend the significance of what’s happening now, right in front of your eyes.

Is this creation or destruction?

The music created with AI above is new. No one has ever heard that music before until now. It is by definition, a pure creation.

And the creation of music is a wonderful thing, particularly original music. Sure, it might lean on music previously, take inspiration from certain musicians, songs and genres but, ultimately, it is original.

Now it’s hard to argue that the creation of music isn’t a wonderful thing. And the same can be said for art, literature and anything from the creative mind.

But does using AI to create something of original nature diminish its beauty? Is it less beautiful because it was created with the assistance of a machine? Does the fact I collaborated with the AI to create that music make it less musical?

You may have a view on that. But when I sent my brother the following tracks, he said, yeah that’s great, but I can also jump onto my computer, open up Ableton (a music-making software) and within five minutes have created my own track not too dissimilar.

He makes a good point. The creation and production of music is easier than it’s ever been. But that’s because of the progress of software like Ableton to help produce it.

If he creates a house track on Ableton, using a computer and software, is it more or less musical than what I just did with the AI? There’s a fair argument to make that it’s the same thing, but an evolution of computing and processing that we already have.

So, on one hand, you need to ask, is AI and the tools we have access to now revolutionary or a massive evolutionary leap forward?

Another idea then comes to mind: if I can create music with AI now, what’s to say in a year’s time that AI won’t be able to do it all by itself? You just jump onto an AI app and ask for the most amazing heavy-rock track ever. Two seconds later, you’ve got it. You then release it and it gets downloads, stream, views, hits – it becomes popular and a revenue-generating piece of music.

Who owns it? Who should own it? And then if AI creates it, where does that leave “real” musicians? Singers, songwriters, instrument-playing musicians? Is the threat of AI so great that creative arts like music are under threat from “the machines”?

So, we also have a philosophical and perhaps ethical question around ownership of creativity and the threat that AI poses to the creative arts.

You can apply and extend this to everything that I mentioned before – art, graphic design, photography, literature, academia… Are these incredible creative industries at risk of destruction due to AI?

Fear vs. optimism

A lot of the current media coverage of AI would suggest that you have everything to fear and nothing to look forward to when it comes to AI.

AI is going to steal white collar jobs, it will eliminate the creative arts, it is an existential threat to the very fabric of society and the extinction of humankind.

Yeah… nah.

In over a decade of research in AI, I’ve concluded that there is absolutely nothing to be fearful of. In all my experiences with AI over my career, the one common thread throughout the development and expansion of what AI can do is that AI is nothing without humanity.

And that if AI were to reach some kind of “consciousness” near the level of a human, it too would realise that it cannot compete or complete anything without us.

I’ve said this before, and I’ll say it again, the future of AI is not man vs. machine. It is man plus machine. Our AI future is one of collaboration with AI, not competition.

This is the crux of the giant leap forward that AI has taken over the last 12 months. That we’ve reached a point with our creation and ability to process data; that the tools we’ve developed that are AI-based now are more effective and useful than they’ve ever previously been.

The two things that previously held back AI were data and processing power. For AI to be effective, it needs to draw on huge data and light speed. The software development was always there to go. But the information was not.

Now it is. And it means we can implement and roll out AI tools in industry and society in a way that will accelerate our ability to innovate, discover and invent.

AI will be one of the most significant economic drivers globally that we’ve seen since the explosion of the “app economy”.

The next big wave of progress and wealth creation will be from the “AI economy”.

Now, that is not a fearful story, is it? That is a story of optimism and excitement. And it’s the truth, the facts, the real story that’s going on right now.

It’s also early in the lifecycle.

I’m of the view that with AI, we’re in a position akin to the mid-2000s and the explosion of the app economy.

YouTube was founded in 2005. Some estimates have its value at over $180 billion to Google.

Airbnb was founded in 2008. Today it’s worth over $82 billion.

Uber was founded in 2009. Today it’s worth over $91 billion.

Facebook (now Meta) was founded in 2004. Today it’s worth over $739 billion.

These are just some of the multi-billion-dollar companies that in the 2000s had only just started. Part of the reason why these companies exploded in value over 20 years can be drawn back to the hardware that enabled their growth: the smartphone…

Or more specifically, the iPhone.

The “iPhone” moment for the like of Uber, Airbnb and the app companies of today was the actual iPhone.

Today’s “iPhone moment” for AI can be drawn back to the release of ChatGPT and now subsequently, Google’s Bard.

The hardware enables the applications. Great software is only great because of the capability of hardware to power it.

The same can be said for AI.

The hardware that underpins all technology is at a level that makes AI worthwhile. And that hardware is only going to get better as we push further along with quantum computing as well.

And that’s where we’re at with the state of play of AI right now.

We have access to the tools (hardware and software) to create the drivers of markets and wealth over the next 20 years. The multi-billion-dollar (and trillion-dollar) companies of tomorrow are only just starting now. Some may not even exist yet.

But you must be open to the idea that it’s now when this wave of the “AI economy” starts and gets underway. Maybe it takes 20 years to realise its full potential, but with the acceleration of efficiency and innovation that AI will allow, maybe that 20-year window is now 15 or 10. And now we’re talking about fast-paced, monumental growth that if you’re caught napping, you do miss out on.

So, what does the future of the “AI economy” look like? What are the key parts of AI that you need to know about and keep an eye and ear out for when it comes to the companies that are just getting started with it today?

The future of AI

There are several exciting areas of AI (hardware and software) worth keeping tabs on, including:

• Generative AI: this is the most hyped area of AI right now. It’s a type of AI system that can create new original, creative content, such as images, text, and music. This has the potential to change how we create, how we interact with computing and how we process and integrate media into our work and lives.
• Federated learning: this type of AI system can learn from data that is stored on different devices. This has the potential to improve the privacy of AI systems and make them more scalable.
• Quantum computing: this type of computing could revolutionise AI by allowing AI systems to process information much faster.
• Robotics and automation: smarter, faster, more efficient robots. Not necessarily the gun-toting kind, but manufacturing and production, agriculture and personal device robots. These go hand in hand with advances in automation that let the robots do things without prompts, but within parameters defined by the user.

It won’t be all flowers and rainbows though. There are several challenges that need to be addressed before AI can reach its full potential. Some of the most important challenges include:

• Data bias: AI systems are only as good as the data that they are trained on. If the data is biased, then the AI system will also be biased. Balance and accuracy of data is everything to efficient and effective AI.
• Explainability: it is important to be able to explain how AI systems make decisions. This is because we need to be able to trust that AI systems are making decisions that are fair and unbiased.
• Security: AI systems are vulnerable to cyber-attacks. It is important to develop secure AI systems that can protect themselves from attack.
• Ethics: we also need to ensure that AI operates in an ethical and responsible way, as well as considered and utilised in an ethical and responsible way. In the same way you see a “celebrity endorsement” labelled as an “ad” today, will you need to label AI music as “AI” or an image as “AI” to maintain transparency?

These challenges also present opportunities as well, so while it’s important to recognise them, it’s also important to look at companies growing from these challenges.

What does this all mean for you?

This month’s headline essay has been all about how you need to look at and think about AI in your world.

We aren’t making a new stock recommendation around AI this month, as we already have one play in our Buy List that does just that – and this month, we want to emphasise that right now, this company is one of the smartest ways to invest in what we see as the growing “AI economy”.

That’s CentralNic Plc (LSE:CNIC) and you can review our full recommendation on that stock right here.

If you want to take a step into AI investing right now, that’s a great way to play it on the London Stock Exchange.

There are, of course, other ways to invest in AI, all over the world. You can do it from a big-picture investment idea perspective, right down to tiny little under-the-radar AI pure-play stocks on markets spanning the globe.

You can be as big picture or as drilled down into it as you like. But for me, one thing is for sure, you absolutely must have exposure to the AI theme and the future AI economy in your portfolio.

You need to have a plan – an attack angle into maximising your potential in this market. It’s high risk, of course, and it’s not for the beginner. That’s why I’ve put together my Four-Step AI Investment Plan, to help people understand AI and then invest in four logical steps.

If you want to learn more about that and gain access to it, head here to find out more.

The AI economy is underway. Fear is not the answer to how this all plays out; opportunity and awareness is key. Whilst this month’s essay is in no way expansive in that it can cover everything that AI will deliver to us, it hopefully helps round out some gaps in knowledge and makes you think about it in the right light.

AI is an area that I’ve been researching for a decade, and I expect to continue researching (and investing in) it for the rest of my life. As such, it will mean that in time, more AI plays on the London Stock Exchange will hit our Buy List.

Make sure to keep an eye out for that, when it comes.

When Sir William Robert Grove, a lawyer turned scientist based in Swansea, Wales, invented the first fuel cell in 1839, he probably didn’t think it would take another 183 years before a company that commercialises the energy-saving technology would turn a profit.

But that’s what has happened.

You see, Germany’s SFC Energy AG, a maker of methanol-powered fuel cells for portable and remote applications, reported in March that it had generated $1.8 million in profit on $78 million in sales in 2022.

According to a recent Canary Media article, that makes SFC Energy officially the world’s first- ever profitable fuel-cell company!

Crack open the champagne!

We jest – though it had seemed like this day would never arrive.

After all, as the Canary Media article states, “publicly traded fuel-cell companies have a long history of losing increasingly large sums of shareholder money”.

“Ballard Power is in its fourth decade as a profitless company. Plug Power lost $724 million on $701 million in sales last year. Bloom Energy’s revenue is growing, but so are its losses,” said Canary Media’s Editorial Director Eric Wesoff in the article posted on 20 June.

As he states, SFC Energy remains the “exception to the rule rather than a harbinger of profitable fuel-cell firms to come”.

So why has it taken so long?

Well, before we answer that, it’s probably useful to provide a quick refresher on the technology, itself.

Fuels cells produce electricity through an electrochemical reaction without any combustion, but they do not need to be recharged like a battery.

A fuel cell can utilise the chemical energy of a variety of fuels, including methanol, ammonia, ethanol and hydrocarbon fuels, to cleanly and efficiently produce electricity.

Fuel cells can charge pretty much anything – buses, trains, military equipment, data centres, warehouse logistics technologies, you name it – all in an environmentally favourable manner.

What’s more, as long as there is a constant source of fuel and oxygen, fuel cells will continue to generate power.

But, despite being around for over a century, fuel cell technology is complex, employing an assortment of electrolytes, catalysts, membranes and temperatures – and it’s here that gives us a clue as to why profits have been so hard to come by.

For the most part, fabricating the membranes is expensive, while the technologies also require costly precious-metal catalysts – typically platinum or palladium – or high-process temperatures.

The input fuels, which range from fossil gas to methanol to hydrogen, can also be expensive.

So, although the costs of fuel cell technology can vary, it’s unlikely that this fully explains why Munich-based SFC Energy – which has also sold more than 55,000 fuel cell solutions since it formed in February 2000 – has now turned a profit when countless other fuel-cell firms never have.

You see, although SFC Energy is active in both hydrogen and methanol fuel cells, it is best known for its work developing the latter, where it is the undisputed world leader in the technology often described as the perfect power generation solutions for a wide range of applications.

“In motor homes and recreational boats, they provide on-board power or electricity in tiny houses,” Bjoern Ledergerber, senior vice-president of Hydrogen and Corporate Development at SFC Energy, told H2 View in March 2022.

Ledergerber added that, “In the industrial sector, they are the green power plant for smart traffic applications, measurement applications for wind farm suitability areas and civil security monitoring at construction sites or outdoor festivals.”

As well as boasting the ability to be integrated into a range of applications, methanol fuel cells are also small, lightweight and easy to install, which helps to lower costs.

In fact, methanol is probably the easiest liquid fuel to make and use in retail settings.

But more importantly, the technology does require a higher input of precious metals compared to hydrogen fuel cells, which can make them more expensive.

And by the laws of physics and thermodynamics, per Joule, if green hydrogen (produced by using electricity from solar or wind) can’t be cheaper than power, methanol can’t then be cheaper than hydrogen.

In fact, according to the Canary Media article, SFC Energy’s profitability has likely more to do with the fact that it operates in a premium-priced niche market – smaller-scale power for outdoor enthusiasts and ruggedized equipment at remote sites – and its reasonable-use case for fuel cells.

And as methanol fuel cells still emit a small amount of carbon dioxide, they do not necessarily offer either a pathway to decarbonization or necessarily profits from the rump of the fuel cell industry.

However, we’re optimistic that other firms will at some point join SFC Energy in turning a profit.

As Canary Media states, ongoing research and development efforts (R&D), some led by the US Department of Energy, are underway to reduce the need for expensive metals and to improve the reliability and lifetime of the fuel-cell stack.

What’s more, in part spurred on by EU mandates and tens of billions of dollars of production tax credits in the US Inflation Reduction Act 2022, fuel cells have the chance to find economic viability, just as wind and solar did over previous decades.

We will keep you informed!