• Generative AI is reshaping the music industry…
  • Graphcore follows quickly after Cerebras…
  • Coinbase to domicile outside of the U.S.?

Dear Reader,

It can be so easy for busy professionals to miss major events these days with so much going on. And no matter where we live, when there is a lot of domestic chaos, it’s those issues that tend to grab most of our attention.

That’s why it would have been easy to have missed major military action in the Taiwan Strait a couple of weeks ago. I have been watching these events very closely considering how critical Taiwan’s production is to the global technology, consumer electronics, and automotive industries.

Mid-April, China conducted three days of military activities targeting Taiwan. Those activities included encircling Taiwan and even practicing ship launched airstrikes from the Eastern side of Taiwan (the side opposite mainland China).

Of course, I’m sure that we’d all like to think it was “no big deal” and these were just routine military readiness drills. But they weren’t. 

This was the second major offensive practice in the last eight months. And to say that China ratcheted up the intensity would be an understatement. To visually understand why, it’s useful for us to take a look at the geography below.

The first thing for us to understand is that Taiwan and mainland China are only about 100 miles apart. And the median line shown above is about halfway between the two. To demonstrate how close that is in terms of military action, a fighter jet traveling at Mach 2 could travel 50 miles in less than five minutes.

What made this military action even more concerning is how it simulated attacks on Taiwan from all directions. China’s forces came from the North, South, directly from the West (closest point), as well as the East from the side of the Philippine Sea.

It was also the first time that China employed an aircraft carrier, the Shandong, in military action concerning Taiwan.

It was a show of force. And a clear demonstration of how quickly China could attack key military targets on Taiwan and bring about a quick capitulation to assert administrative control over Taiwan.

China’s actions around Taiwan unsurprisingly prompted concerns from Japan. Given the close proximity to Taiwan, Japan scrambled its own jets after monitoring takeoffs and landings of about 80 fighter jets from the Shandong.

When living and working in Japan, we were never really worried that there was any threat that China would attempt to attack Japan. The primary concern was one more of global geopolitical stability and international commerce. Japan has fond ties with Taiwan and is also closely intertwined with Taiwan’s semiconductor and consumer electronics industries.

And Japan has a deep contextual understanding of what is going on right now. It has historical and cultural context that most of the West simply doesn’t understand. Japan knows that times of crisis provide a catalyst for things to happen. It’s literally written into the language.

In Japanese, the word for crisis is 危機, pronounced “kiki.” The first of the two kanjis mean “dangerous,” and the second kanji can mean “opportunity.” Some of us may have heard the refrain, a crisis can mean both danger and also create opportunity.

This meaning is perfect for the current situation. With all the geopolitical tensions, the regional and domestic issues in the U.S. and Europe, and non-governmental organizations like the WEF causing chaos and upheaval with its well-funded agendas, I doubt China will have a better time to make its move than while the Western world is distracted with all of its own problems.

There seems to be little interest from the West in getting involved in yet another conflict. That’s why China most certainly sees a great opportunity. And I believe that we’ve just seen a sign of things to come in the very near future.

Is that a new song by Drake? Or an AI?

Have you heard the hit new song?

It’s called “Heart on My Sleeve,” by the popular artist Drake… sort of.

In reality, somebody going by the pen name “Ghostwriter” used generative AI to produce a new Drake song.

And get this – Drake’s fans loved “Heart on My Sleeve.” It’s now been played well over 15 million times. Those listening to it would never think that the song was completely AI-generated if they didn’t know already. Most can’t actually tell.

Of course the big question now is whether the music industry will go after Ghostwriter for creating a song using Drake’s name, image, and likeness. It’s certainly illegal to profit off of somebody else’s “intellectual property” without their consent.

Bigger picture, this gives us a glimpse of what the music industry will look like going forward.

Imagine what happens when the industry starts training generative AI on an artist’s entire body of work. Suddenly the AI could start cranking out new songs continuously. In fact, it has already happened. An entire eight track album made in the likeness of band Oasis was just created by an AI… An entire album. And it sounds great.

And the remarkable thing is, the artists wouldn’t even need to sing the songs. They would simply need to approve of what the AI produced for them.

So this could be an incredible tool for the music industry to multiply its productivity in a big way. We could see an onslaught of AI-generated material from popular artists in the months to come.

Music companies will have a large impact on the direction of this trend. I expect that they are already incorporating new language into contracts that stipulates the right to produce and market new music in the band’s name by using generative AI technology.

Bigger picture, at some point the industry wouldn’t even need human artists anymore. They could create an AI-generated persona and market it just as they do human artists.

That may sound odd at first. But it makes perfect sense.

After all, the major music labels often have problems with their top talent. Because the talent is human. They go through their ups and downs, drug and alcohol problems, and personal issues. Sometimes they do fantastic work and are productive… Other times it’s a disaster.

With AI-generated music, the music labels could potentially move away from human talent all together. They could create everything using AI… and then they wouldn’t need to pay out any salaries or royalties to the artist. Their profitability would skyrocket. The downside of course is not being able to go on tour with an AI. And touring, in person interviews, videos, etc. that require a human “star” are critical to the business and very lucrative for the music labels. 

This concept isn’t just limited to music. We should expect the same developments to unfold in film and television. It’s possible that – sometime in the near future – one of the most popular actors will be an AI-generated character.

This is another example of how generative AI is transforming the world in ways very few expect but most will enjoy. Who wouldn’t want to hear a new album release from a long since passed musician, or a new movie from a favorite actor who has long since shuffled off their mortal coil?

Graphcore takes things one step further with its AI…

Speaking of generative AI, a huge development just caught my eye.

AI-semiconductor company Graphcore just used its chips to train and release a new large language model called Dolly 2.0. It’s a generative AI that functions very much like OpenAI’s ChatGPT.

This alone is certainly an interesting development. As we discussed last month, generative AI trained on AI-specific semiconductors is one of the fastest growing trends in the industry.

And as long-time readers remember, Graphcore’s semiconductors are quite powerful. The company calls them intelligence processing units (IPUs) because they are designed exclusively for AI applications.

Here’s what the IPUs look like:

Source: Graphcore

As we can see, Graphcore’s AI hardware is highly sophisticated.

But what’s most interesting here is that Graphcore did what nobody else in the industry has done yet – it open-sourced Dolly 2.0 and made it available for commercial use.

That means anybody can take the code, modify it, and then rebrand the modified generative AI for their own applications. Graphcore will not retain any intellectual property rights to modified versions of Dolly 2.0. And whoever uses the underlying technology and generates revenue from it gets to keep all the profits.

This move allows any organization or individual to quickly create their own customized generative AI with a similar functionality to ChatGPT. Except each organization can train the AI on data that’s customized for its own applications.

Basically, Graphcore has removed the economic barriers to accessing and building a business based on generative AI.

Of course, this begs the question – why would Graphcore do this? Why give away its tech?

The answer is that Dolly 2.0 runs best on Graphcore’s semiconductors. So Graphcore’s strategy is to seed the market with the AI software in order to sell more of its hardware. This is the same strategy that NVIDIA used so well over the years, and it’s also the same thing that Cerebras just announced a couple weeks ago.

Graphcore’s hoping to cash in on the generative AI boom with this move. It’s a reasonable strategy, and it is a precursor of things to come.

My longtime readers will remember that semiconductor companies were one of the biggest beneficiaries for the mass adoption of cloud computing in the 2010s. Companies like NVIDIA soared from under $10 per share when I recommended it in a closed-door presentation in 2016 to $286 today. That’s a massive gain of over 2,700%.

So we’re going to keep a close eye on semiconductor companies that are servicing the generative AI space. The scale and speed of adoption for generative AI could rival the cloud computing boom we saw 10 years ago. And some of the best investments in this space could very well be the companies providing the necessary components.

The SEC is running Coinbase out of town…

We’ll wrap up today with a disconcerting development.

Speaking at a financial technology (fintech) conference in London, Coinbase CEO Brian Armstrong told attendees that Coinbase may relocate its headquarters outside of the United States. That’s a direct indication of just how hostile the regulatory climate is towards the digital assets industry in the U.S. right now.

What’s ironic here is that Coinbase is one of the most buttoned-up digital asset companies out there. The firm made regulatory compliance a top priority from the very start. This is one of the reasons I invested in Coinbase as an angel investor all the way back in 2017. And it’s one of the very few companies in the industry that was willing to undergo scrutiny by the Securities and Exchange Commission (SEC) to go public.

As such, Coinbase operations are transparent through its quarterly and annual filings with the SEC. Anyone can look at its quarterly filings and see how well, or not, the company is being run.

Yet, the Securities and Exchange Commission notified Coinbase that it’s likely to pursue a lawsuit against the company through the issuance of a Wells Notice. Despite the fact that Coinbase hasn’t done anything improper or outside the scope of its business activities outlined in Coinbase’s original IPO filing that the SEC approved.

So it appears everything is coming to a head in the digital assets space right now.

As for its most likely move, the United Kingdom makes a lot of sense for Coinbase. The company already has about 300 employees in the U.K. And it also has regulatory approval as an e-money provider in the country.

Of course, if Coinbase leaves the U.S. it will be taking a lot of jobs and investment capital with it. That would be a bad omen for the future of the digital assets industry in America.

In many ways, Coinbase has been an “anchor” for the digital assets industry. It’s a fully compliant, user-friendly platform. And it’s the only publicly-traded digital asset exchange to date. That went a long way to help legitimize the industry in the eyes of traditional finance.

Two more years of antagonistic policies could irreparably damage the blockchain industry in the U.S. But it won’t stop the amazing development that is happening elsewhere in the world. It will simply put U.S. businesses and labor force way behind in the industry in investment and knowledge around the next generation of internet technology.


Jeff Brown
Editor, The Bleeding Edge