Peak attention

We’ve reached peak attention on the internet. As you can see from the graph below, the time spent online is not significantly growing anymore.

The growth of the internet has generated 4 champions (Google, Apple, Facebook & Amazon) who together dominate for the most part how you spent your time online, where you do your shopping and how you get online.

Comscore’s Mobile App Report 2017 underlines this, they found that most of all our time we spent only on 10 apps.

And those apps are:

These companies are now at their peak and they have the momentum to buy, absorb or change their tactics to fend off any competitor. The Facecbook / Snapchat battle is a great example of that. Facebook just absorbs everything which makes Snapchat unique. It might not capture all the audience away from Snapchat, but just enough to defend their position.

Distribution on the internet is becoming harder & expensive

VCs love to say that it became much cheaper to start a company and in many ways this still holds true today. Except it became more expensive and harder to distribute your product – being it a service, software or product – and reach your audience.

It is more expensive because today you need a website and 2 mobile apps instead of just a website.

Harder because by default you don’t have easy access to your audience anymore. The audience does not live on the web anymore. They live in Facebook, app stores, search engines, Amazon and Netflix. The attention of the average internet user is fully engaged and it has become virtually impossible to steal a bit of that attention away without smart, engaging and costly marketing.

You used to be able to start a company by building a site and service and launching it. You could market it to a few key users and find a few journalists to write about it. Today that is not enough, you get quickly buried because we reached peak attention.

I would argue that it is much cheaper to scale your company nowadays – not necessarily to start one.

I don’t expect that AI, machine learning, AR or VR will not significantly change the landscape. GAFA is well situated to absorb and embrace those new technologies. They’ll lead to new opportunities but companies need to battle the same attention problem like any other company launching a new product or service on the internet.

It’s reasonable to assume that only a major new platform or major platform shift can threaten their position and generate a new set of GAFA companies.

Web vs native apps

The big promise of web 2.0 was that eventually all applications would run inside a web browser and that native apps would go away. This was in early 2000s. We’ve come a long way since then but native apps have not gone away – even by a long shot. Especially the birth and rise of mobile threw a wrench in this vision, but there are some early signs on the horizon this is going to change.

The IOS and Android mobile platforms have been taking over the world. 51.3% of internet traffic originates from a mobile platform and the developing part of the world depends even more heavily on mobile platforms. But at the same time, mobile platforms have been killing the web. 60% of mobile usage is inside apps nowadays and it’s growing. Apps are different because they are built using lower level programming languages like Objective-C (IOS) and Java (Android). Because of their low-level nature, they’re more costly to develop for plus two different versions need to be maintained to support both platforms. Luckily – you could argue because of this – we only ended up with two mobile platforms.

The reason the web took off the way it did was because it could be accessed from any device and OS. A web browser was enough. This allowed for websites to quickly scale because the distribution via web was easy, fast and enormous.

But the web – an open platform governed by open standard bodies – was not ready to run on underpowered mobile platforms. The same applied to mobile networks at the time. When the iPhone came out in 2007, it supported AT&Ts 2G EDGE network with a maximum transfer rate of 200kbps. The first iPhone had a battery-friendly downclocked Samsung ARM CPU which was 100x slower than today’s iPhone 7 and IOS Safari’s web performance increased by 200x. Not only the mobile hardware improved significantly over time, but also the mobile networks are significantly faster to the point that in many places of the world there’s no significant speed difference between fixed and wireless connections.

In the early years of the iPhone, Steve Jobs was bullish on the future of the web evidenced from his controversial Thoughts on Flash open letter to Adobe. Early on IOS browser Safari came with features out of the box to bring web sites into IOS by allowing them to run full-screen and give them a place on the home screen with an icon like any other (native) app. These features still exist today.

But Apple and the rest of the world quickly realized that to get the necessary performance, 3D acceleration, grant access to local features like the camera, GPS and storage, it was necessary to move to native apps. Apple introduced their IOS app store in 2008. Apple’s App Store is an enormous success and at the same time moved the focus away from the web as a viable app platform.

W3C, ECMA and others kept chugging along and kept moving the web standards forward. We are now at the 5th standard for HTML, 7th edition of the JS/ECMAScript standard, 4th level of CSS and version 2 of WebGL. Javascript was famously created in 10 days by Brendan Eich because Marc Andreessen wanted to augment the web by giving their browser some limited dynamic capabilities and it needed to be done before the Netscape 1.1 release. This happened in 1995.
The web has come a long way since then and now has 3D support, local storage, notifications and access to location services.

If we would imagine a world where a mobile computing devices has access to unlimited processing power, unlimited storage and unlimited network speeds, there is certain a great case to be made that the web will come back as the platform for apps. As shown earlier, we are moving towards a world like where the resource constraints – which made native applications become the default choice – are no longer an issue.

It stands to reason that the web will become as prevalent as it has become on the desktop and arguable even bigger than it is today. In 2009 Apache Cordova – formerly known as NItobi and PhoneGap – was born. Cordova was one of the first tries to work towards unifying mobile platforms using the web, but is no longer alone in morphing the web into native app development. In 2013, the Ionic and Electron frameworks came to life. All of these open source projects have the same goal which is to marry web standards (JS, CSS, HTML, WebGL) with native underlying platforms breaking free from the constraints of the generic web browser. They tie web technologies with a native experience on each supported platform while allowing the same application without modification live on both platforms.

The benefits are clear. A single app which can run on any platform with little to no modifications built using a higher level programming language is going to make it cheaper, faster and easier to build and distribute new apps.

The problem of native app development is exacerbated by the large number of display sizes in the market today. Apple’s IOS lineup from iPhone SE through iPad Pro there are 6 different display sizes to support. Android is in a similar situation. The larger tablets move into desktop territory.

Talking about desktop territory, Google’s strategy to bring Android to ChromeOS laptop feels like an impatient move to make that computing platform more relevant. In many ways ChromeOS was meant to do the same by graduating web applications to become standalone applications on a computing platform. Much of the strategy to bring Android applications to ChromeOS stems from the positioning of Chromebooks as lower-end alternatives to their Windows and MacOS counterparts. It’s unclear to me if that was a conscious decision by Google or the market – namely OEMs designing and manufacturing Chromebooks – decided to position chromebooks that way. Though Google has been very successful in marketing their Chromebooks in the education market.

Below the surface there’s an additional opportunity. When it becomes easier to develop for multiple platforms simultaneously, it also creates opportunities for new platforms to emerge. For instance, Microsoft has been struggling to get a foothold in the mobile market and much of that can be attributed to their slow response to the mobile computing platform. When they were moving, it was already too late. The native app ecosystem proved an enormous hurdle to cross and app developers were not waiting to support another platform. But there’s a possible future where that hurdle is going away and that is an enormous opportunity for Microsoft and others.

There’s a fair chance that we look back on the era of the native mobile apps as abomination of the norm.

Future platforms

Google, Apple, Facebook and Amazon (GAFA) are now the most valuable and successful companies in the world. This also means that they’re maturing rapidly and this means the platforms they ride on are mature quickly as well. These companies are carrying boatloads of money which give them the power of money to compete against any new (potential) threat quickly. A good example is Google Home and the subsequent push by Google to bring this product to as many consumers as fast possible. It is competing against the highly successful Amazon Echo family of products. 

Another example is the introduction of “snapchat” features into the different properties of Facebook. Instagram and WhatsApp reached 200M DAUs within months – just compare that to Snapchat’s 160M DAUs on IPO which took them years to get to that point.

So strictly speaking any new development on the internet or mobile will either be quickly bought up or copied by these companies. And this will make it much harder to launch and grow a successful competing company on these platforms.

History also tells us that when companies reach this height and as their platforms further mature, new platforms and opportunities will arise. And interesting enough these companies will not be able to compete or turn around their business to take full advantage of this. This might sound inconceivable today, but let’s look at these two examples. 

First, there’s Walmart, the biggest retailer in the world, but in online retail they’re irrelevant. You’d think that a company with the resources and knowledge of retail would be able to compete and keep up with online retail competitors. But they are not. Of course, they won’t go away, but they will not become the number one online retailer anymore. 

The same can be said about Microsoft. They completely missed the boat on the internet and mobile platforms until it was too late. You could even argue they missed the cloud platform as well. Same applies to Intel. Intel missed the boat on mobile as well. Together they were known as Wintel and they were the jewels of the nineties. Now, these companies have become irrelevant from a mindshare and technology direction perspective.

Before Wintel, there was IBM. IBM was the computing company of the 80s, but today they’re irrelevant.

Of course IBM, Microsoft, Intel and Walmart are still alive and generating sizable revenues. Irrelevancy does not mean bankruptcy – in many ways, these companies are too big to fail. But they’re on the infinite road of playing catch up in their respective markets.

This brings me to my point that the hegemony of GAFA will go away as well. You already see some early signs of this in both Google and Apple. Google spends an extraordinary amount of money ($15B in 2016) on R&D but no new platform has emerged from it yet. Same applies to Apple. Their biggest product launches in the last 5 years were just accessories – watch and AirPods – and not platforms. They’ll never be able to grow into revenue streams similar to iPhone or Macbooks.

Taken together, you could come to the conclusion that we’re on the brink of a new major platform to emerge. Investment dollars and ideas are turning away from more traditional investments in internet services and mobile towards other platforms. At this point, it is impossible to say what these new platforms will be. It is pretty certain that they will emerge from all the attention towards blockchains, machine learning/AI and augmented reality/virtual reality. None of these technologies is a platform but you can bet on it that it will be enabled by one or more of them.

Platforms emerge when there is a technology breakthrough which makes it possible to bring technology to form factors or places which were previously impossible. 

Electric autonomous vehicles are way on top of my list. It will be as disruptive as the internet. Cities will change; no need for parking lots, people can live further out in suburbia. Car-ownership will change, there might not even be a need to own a car anymore. This will impact car dealerships and car manufacturers. We might only end up with a few of manufacturers and reach complete commoditization of car manufacturing. There will be no further need of gas stations. Cars are probably maintained by large corporations running major fleets of autonomous vehicles. The car wash will go away. There is no longer need for the DMV and maybe policing of traffic is a thing of the past. Tax revenue streams need to be readjusted.

A little further out but as fundamental as the computer processor or the internet, I would bet on either power consumption or power storage technology. The biggest problem holding back many potential applications, form factors, and new technology is power consumption and power storage. Any new technology which can decrease power consumption or increase power storage densities by 100x is going to change the world fundamentally.

It is not limited to power consumption or power storage. The new technology can also change the paradigm on how devices work. For instance, power could be transmitted wirelessly. Or the technology can offer a way to offload 99% of processing to a different device while offering a full-fledged platform performance.

When that happens it is pretty obvious what is going to happen. History is telling us that again. Typically an emergence of a new technology is unbundling existing platforms and then bundles it again into the new platform. For instance, the internet unbundled many things like the cable subscription, magazine subscriptions, music distribution, phone subscriptions and this is still going on. Currently, a lot of offerings of the banking industry is being unbundled like debit cards, loans, mortgages, and insurance. But at the same time, you see new bundles appear. GAFA is a good example of this. They bundle and protect their own platforms. The openness of the web has been replaced with proprietary – and much less open – platforms like IOS and Android. The mobile phone bundles everything like your wallet, camera and must player. Facebook is pulling everyone into their walled garden much like AOL did at the time with their own version of the web.

And this is my second major observation for the next platform which is openness. All the major platform shifts were twofold. First, the emergence of new technology which allowed new form factors, services or applications not previously possible before. And second, the emergence of an open standard to make us of the new platform. It is essential to level the playing field for a new technology to become successful. For instance, if the new technology would be of the nature of wireless power transmission, it would only start to take off when the platform or protocol is open.

When this all happens the whole cycle will start over again.

Artificial scarcity in the broadband market

Last week the Wall Street Journal wrote about the introduction of data caps by cable companies. Comcast, AT&T and many other cable broadband providers apply data caps or are experimenting with putting data caps on their connections.

Here are few quotes from the article:

Comcast technically has a 250 gigabyte monthly limit on its 23 million Internet customers but stopped enforcing it in 2012. The company is running a series of trials with a data threshold of 300 gigabytes, and, in some areas, varying thresholds and an unlimited option for an extra fee.

Until recently, all of AT&T Inc.’s broadband offerings had limits ranging from 150 gigabytes to 1,000 gigabytes depending on a home’s connection speed.

For limited plans, AT&T charges $10 for every 50 gigabytes over the limit; Comcast charges the same.

Time Warner Cable says the company’s average household usage in December was 141 gigabytes a month and has grown about 40% a year.

Comcast says its aim is to ensure the heaviest users are paying more than lighter ones, since 50% of its bandwidth is consumed by just 10% of its customers.

AT&T says the change will include offering unlimited data to people who pay an extra $30 a month or who subscribe to DirecTV, which it owns, or its U-verse television service.

This is typical cable company behavior. In many areas, these companies enjoy a monopoly — or in the best cases a oligopoly — on their services. For this reason, these companies are completely focussed on extracting as much as they possibly can from their customers. Since there is no competition to speak of they can charge whatever they want. The pricing game for them is purely about maximization of revenue and ultimately gross margins.

However, the market is changing. Consumers — and especially the younger demographics — are ‘cord cutting’. The cable companies made lots of money with offering Triple Play packages (Internet, TV and telephony) but now they are confronted with changing behavior which leads to price erosion. People who choose to get internet only packages bring in less revenue.

Because of price erosion and declining revenue from their cable subscriptions, these companies need to find a way to combat this. The data caps are introduced for exactly this reason. It’s their way of ‘easing’ their customers to get back to paying more.

Cable companies know exactly how much money they can extract from their customers. Now they find themselves in a situation where that is not set in stone anymore. They’re rethinking how they price and sell their services. They are introducing artificial scarcity in the market to stop the price erosion.

In the worst possible scenario this means we end up paying more for all services. Cable companies will get us back to Triple Play pricing over time and now we will pay companies Netflix, Hulu and HBO separately for their streaming services.

‘Net neutrality’ how lofty in its goals will not bring us anything to prevent that. Ultimately there’s a total lack of competition in the internet broadband market and as long that’s the case, we will end up paying more for less.

This quote from The Verge sums it all up:

So schemes like data caps, which have already been used extensively in the wireless industry to reap as much money as possible from customers, exist solely to frustrate those customers — which is really an incredible situation in a country that ostensibly cares so much about the virtue of competition.

The post-finishing waste land

When you show a 3D printed product to someone who has not seen a 3D printed piece before, there is significant chance that the conversation will be about the material. And that is not surprising. 3D printed pieces look rough, show “printing lines”, and feel different than regular materials.

For purely functional pieces that is not an issue. But often the aesthetic component is as important as the functional component of a product.

It is not so much about the capabilities of what can come out of 3D printer. It has been proven that a lot of things can be 3D printed. But is the material of the right type to be acceptable for a particular use case.

Does it have the right weight?
Does it have the right feel?
Does it have the right texture?

We can wait until 3D printing processes improve with new materials and new deposition methods to overcome the quirky aesthetic. But at the same time, there are real options out there to apply existing finishing options to 3D printed pieces.

The challenge is that many of the post-finishing processes are geared towards mass-production of products. And this does not apply to 3D printing. I have this beautiful vapor-smoothed FDM printed bowl. It is gorgeous but the process never made it to mass 3D printing because it was too cumbersome to scale. The interesting aspect of it though is that whenever I show it someone nobody asks about the material anymore.

I see many companies struggling with the concept of bring 3D printed parts to live with post-finishing. It brings a whole new level of complexity. And it needs more attention.

Download a human

Copy. Download. Live Forever.

People who are close to me know I am a Star Trek fan. I also have a long time wish to go to space — maybe some day. In my home, I have a iPad dedicated to the HD Earth Viewing Experiment on UStream which sends out a constant video feed of three cameras attached to the international space station ISS.

But secretly I do not believe in a Star Trekian future. Seriously I do not think it is a plausible future that we space travel as humans beyond our solar system. I think even Mars is quite a stretch.

Instead I think it makes much more sense to download our human brain into a computer and send that computer instead. Just imagine there is no need for food, water or other supplies + equipment necessary to make those journeys. It even does not matter so much how long the journey will take.

Computing power is progressing nicely and IBM already designed a computer to simulate a mouse brain. In another decade or so and we will see computers who are capable of emulating a human brain. Now at that point we can program an artificial intelligence which is indistinguishable for a human. We can program the computer anyway we like, but why not model it by copying actual humans? You might think I am insane, but it would make so much more sense to me.

The hardest thing is giving away

You always end up with less when working at an early-stage startup

People working in startups are all so familiar with the concept of giving away responsibilities. Whenever you start in a startup, you can count on the fact that you’re role is changing all the time. Regardless of whether responsibilities shift downwards or sideways, they are inevitable moving somewhere.

Your startup starts out with 3–4 people who do everything. Slowly, new people are added, roles are carved out, and responsibilities are limited. With each new influx of people these roles are reevaluated and the same ritual starts again.

The same applies as the company grows your personal capabilities are constantly evaluated. Your position within the company is constantly shifting and changing. The company often grows faster than your personal growth can keep up with. Or your role is carved up in 2 or 3 different roles and you have to choose between them.

All these things have nothing to do with failure or underperformance. It is about making the company grow as fast as possible while keeping everything manageable. But I can tell from experience it is sometimes hard to let go. Even it is going downwards instead of sideways. You might be end-responsible, but someone else is actually executing it. It is not the same.

It is for this reason that startups are not for everyone. You need to adapt and be comfortable with adapting. You have to embrace change and you need to be comfortable with finding your own limitations. But more so, nothing better to run into this because of the company you work is successful and growing up. On the end of the day, there is nothing more satisfying than bringing something to the world which is self-sustaining and impacting the world in its own way.

Nobody cares about customization

Yes please on size, fit, material and color. No thank you on everything else.

With Shapeways I had a lot of interest into customization. It would make the perfect use case for 3D printing. I have been following the developments in the industry for some time now. But up until today, I can only conclude that people are not so interested in customization.

I think it is because people do not want to think about design as much as we like to think. They rather leave it up to the designer to design what they want.

Now this does not mean all customization. I primarily talking about customization with regard to the design of an item. There are plenty of opinions on color, size or fit. Just take a look at anybody’s iPhone and you see plenty of different background images and different arrangements of the apps on the screen. Also, people like to choose the color of their iPhone.

It is like buying a couch. I have an opinion on the size of the couch — two-seat, love-seat or three-seat — and the color. I have a preference for the material. But I never have had the desire to change the design. I just choose the one I like.

Same goes for clothes. I love to be able to select on size — it should fit right? — or the color. But I would not want to design it myself. I would like an idiot. I am not a clothes designer and never aspired to become one.

In the end it is about offering customization options which are relevant to people. This could be far reaching — from fit, color, size to material — but it almost never touches the design itself. The design is the domain of the designer. Everybody is fine to keep it that way.

Where we manufacture and the impact of 3D printing

In my series on the impact of 3D printing, I wrote about my views of the impact on supply chain and product design. In today’s post, I write about manufacturing locations. This topic has already been partly covered in the post about supply chain, but I think there is more to say about it.

Manufacturing takes places all over the world. But the bulk of our manufacturing takes place in Asia. I expect that 3D printing has the opportunity to change that. 3D printing offers benefits over mass production like:

  • one-off and small series production
  • multiple products coming out of the same production line
  • semi automated production lines
  • smaller factory footprint because of all the above.

Looking at these benefits, I see new directions for manufacturing, which have an impact on the location where production will take place. The directions I see are:

  • manufacturing closer to major markets — in essence bring manufacturing back to the West localized production near large population concentrations
  • insourcing of production by non manufacturing companies
  • around the corner production (Kinko concept)
  • home manufacturing (home printers)

Manufacturing closer to major markets
We produce in Asia because it is cheaper to manufacture overseas and ship it back to us than to produce it locally. Now this concept works great for mass-produced goods, but with 3D printing it may change. With lowering prices of 3D printing equipment and materials, it can tip the balance and make localized production possible.

Localized production near large population concentrations
Given the above, it makes more sense to move production closer to large cities or near central distribution hubs similar to warehouses of big retailers like Amazon. I can even imagine that some production will take place inside those existing warehouses.

Insourcing of production by non manufacturing companies
Beyond making production local, certain companies or organization can start producing themselves. The benefits to be able to produce on demand on-site can in certain cases very compelling. For example, a hospital which needs specialized tools for operations or tailor-made items for patients. Or car shops who need to special parts to repair a car. In these cases, it saves time, costs and inventory to have produce them in house.

Around-the-corner production (Kinko concept)
When 3D printers become more capable and their use more ubiquitous, the next step is to go hyper local. Local shops in convenient locations (a la Kinkos) open and offer local pickup and on-demand 3D printing options to businesses and consumer alike.

Home manufacturing (home printers)
And there is the option of home printers where consumers can print their own products at home.

There are considerations to take into account on how the future will unfolds itself. One consideration is the acceptance of customers of build-to-order products versus off-the-shelve products. Build-to-order offers freedom, but at the same time does not deliver instant gratification. Off-the-shelve delivers that instant gratification, but the customer is limited in choices. The other consideration is that the applicability of 3D printing is different in each product category. The applicability could be limited to shells or components for some, while others are completely manufactured using 3D printing.

There are many opportunities for 3D printing to have a major impact on the manufacturing locations of products. If you look at 2D printing business environment, I can imagine something similar for 3D printing. In 2D printing, you have many options to print. Each of these options is specialized in certain markets or services. But foremost they are complementary to each other. I expect nothing less for 3D printing.

Impact of 3D printing on supply chain

When 3D printing becomes mainstream, it will have a major impact on many aspects of manufacturing and design processes. In the next weeks, I am writing a series of posts on how I think the 3D printing revolution will unfold and what impact it can have. In this first post, I am writing about the impact on supply chain.

So what is supply chain? Wikipedia defines it as:

A supply chain is a system of organizations, people, technology, activities, information and resources involved in moving a product or service from supplier to customer. Supply chain activities transform natural resources, raw materials and components into a finished product that is delivered to the end customer. In sophisticated supply chain systems, used products may re-enter the supply chain at any point where residual value is recyclable. Supply chains link value chains. (source: Wikipedia)

A typical supply chain for a product goes from product design to delivery to the customer. In the following graph is a simplified example of typical supply chain for a product. As many products, it is manufactured in Asia, and sold in one of the Western countries.

Now let us assume that 3D printing and personal fabrication become mainstream — either using a personal 3D printer or a local service. How does the same supply chain look like then?

As you can see, the supply chain is greatly simplified. The simplification is possible, because of a unique characteristic of 3D printing, which is the ability to manufacture different designs without building a complete production line for each and every product. It opens up the doors for localized production facilities close to consumer concentrations.

Another benefit is that there is no need for the production of large batches to offset the investments necessary for setting up the supply chain, tooling, production lines and transporting the product to the customer location. This reduces risks in the supply chain for product failures. With 3D printing, the production can take place in small batches or as one-offs, which makes it possible to adapt a product design almost immediately.

To summarize, the simplification of supply chain using 3D printing can lead to:

  • Shorter lead times
  • Reduced supply chain risk
  • Reduction of transportation costs

Today 3D printed parts are still expensive compared to their mass-produced counterparts. But I am confident that gap will close in the coming years. The growth of the industry will lead to lower prices for machine and materials. Another caveat is that not every product can be 3D printed yet — either as a whole in one go or even in parts. That is a bigger gap to close. Though options exist to mix-and-match products sourcing both from 3D printed parts and off-the-shelf mass-produced parts. I expect that the transition will be more gradual than revolutionary.

I see significant benefits for 3D printing and personal fabrication for supply chain. It is not surprising that high-end manufacturers like Boeing, EADS and General Electric are already actively using or researching the options for 3D printing for the manufacturing of their products. As 3D printing further matures, the technology becomes within reach of a larger group of manufacturers. At some point, the manufacturers like we know today will vanish. Just like they did when all manufacturing moved to Asia.

Endless forms of iterative design

A while ago I wrote about Iterative Design and the potential of iterative design on how products are created in the future. Now Hod Lipson and his team brought this concept to the next level with their EndlessForms website.

They combine evolutionary algorithms and generative encodings with crowd sourcing of designs. The results are interesting 3D designs. You could argue the usefulness of many of the designs on the site but it does show a powerful new and easy-to-use way to 3D design and iterative design.

How does it work?

The user selects one of created designs in the gallery or chooses to start anew. The site then evolves this design in 15 different directions and shows the user 15 options. The user can then choose to explore one of the evolutions further, combine multiple evolutions and explore that, or save the 3D design. Optionally the user can choose to 3D printed his design — which is disabled at this moment for some reason.

Here is video on how it works:

Their inspiration was Charles Darwin and the evolution of species. They found their name in the Origin of Species.

“… from so simple a beginning endless forms most beautiful and most wonderful have been, and are being evolved.”
— Charles Darwin, On the Origin of Species

I love these types of innovations for new ways to create 3D designs. It makes 3D design accessible and fun.

Why AI is hard

Artificial intelligence was one of the first buzzwords I can remember from the previous century. It promised a future with intelligent computers or devices which could understand you and act autonomously.

Up until now we still do not use AI-enabled devices in our daily life. Why is that?

Our brain has 10 billion brain cells. Each brain cell has 10.000 connections to other brain cells. Neurons activate pathways to brain cells. The activation of brain cells along those path pass along the message of the neuron to other brain cells. Some pathways end and the message is not forwarded. The combination of brain cells and pathways between brain cells is how the brain processes and stores information.

To make AI happen we should be at least be able to mimic the brain cells and pathways functions. There are two options available; hardware and software emulation.

You could compare the brain (without memory) with a CPU. Current top-of-the-line computer processors contain 2–3 billion transistors. That comes close but each transistor has only a few pathways and is in the end only a transistor. If you would emulate the pathways the current CPU need to contain at least 50.000 more transistors than they today. Based on the Moore’s law (CPU transistor count doubles every 2 years) we can have to wait until 2025 to get to this point.

From a memory perspective you need to store the values of each of the pathways of the brain cells. This would require 10B * 10k * 1 byte = ~90TB of data. The biggest individual memory units currently available are 12GB. Applying the law of Moore again and we reach 90TB memory units in 2023.

With current technology we are able to emulate the brain function of an insect. Hence you see all those insect-like autonomous robots.

Based on these calculations it looks major progress can be made in AI in 15 years from now because the necessary computer hardware is available to be able to emulate the brain functions. Of course emulation is only the first step. The emulated brain also needs to be programmed to work. Much like any brain is programmed before birth.

Median pay of S&P500 companies

The Wall Street Journal has a interesting article on median pay of S&P 500 companies which contains a few of the tech giants. Here are some interesting numbers:

  • S&P500 median pay: $69k
  • Alphabet / Google: $197k
  • Facebook: $240k
  • Amazon: $28k

Obviously, the composition of the work force of each of these companies greatly influences the median pay. Interestingly enough Apple is one of the companies missing from this report even though it’s part of the S&P500 AFAIK.

Do not disturb

I came across this article dumber phone and it expresses well about how I feel about my computing devices today.

Smartphones are useful, but they are also incredibly addictive, and that addiction is at the epicenter of Silicon Valley’s effort to grab an ever increasing percentage of our minds. In that a substance or behavior controls us, it becomes our master, and that just won’t do.

Amen. There’s a lot of good advise in this post.  I’ve been battling this very thing for a long time. Not only on my phone, but on all my devices. I carefully curate notifications. By default notification are not allowed and most applications do not get notification privileges. Notifications are off for all apps on my laptop and iPad. My laptop is for working and my iPad is a distraction-free consumption device. I only enabled calling, text and email notifications on my phone.  

Even websites want to send notifications nowadays. I can’t even phantom why you would want that.

Apparently Nir Eyal wrote a whole book about it and then some which is ironic since he’s well-known for his other book; Hooked: How to Build Habit-Forming Products. Perhaps it takes one to know one.

The Kamprad Test and Moravec’s Paradox

An article in the Economist reports on efforts of AI researchers in Singapore to automatically assemble an IKEA chair. They call it the Kamprad test – named after the founder of IKEA:

COMPUTERS have already proved better than people at playing chess and diagnosing diseases. But now a group of artificial-intelligence researchers in Singapore have managed to teach industrial robots to assemble an IKEA chair

It took a pair of them, pre-programmed by humans, more than 20 minutes to assemble a chair that a person could knock together in a fraction of the time.

Clearly Moravec’s paradox is that play here. The paradox is:

Moravec’s paradox is the discovery by artificial intelligence and robotics researchers that, contrary to traditional assumptions, high-level reasoning requires very little computation, but low-level sensorimotor skills require enormous computational resources.

It can even explain the difficulties Elon Musk and Tesla face to fully automate their production line.


Unbundling the internet and data ownership

I was reading the two articles below today. Their observations are correct, but their arguments are very technology driven. Technology is just an enabler, it’s ultimately successful business models which drive innovation and change.

The Missing Building Blocks of the Web

The web was designed so that everybody was supposed to have their own website, at its own address.  

An individual having a substantial website (not just a one-page placeholder) is pretty unusual these days unless they’re a Social Media Expert or somebody with a book to sell.

There are social barriers, of course — if we stubbornly used our own websites right now, none of our family or friends would see our stuff.

It’s time to rebuild the web

I’ve written several times (and will no doubt write more) about rebuilding the internet, but I’ve generally assumed the rebuild will need peer-to-peer technologies. Those technologies are inherently much more complex than anything Dash proposes. While many of the technologies I’d use already exist, rebuilding the web around blockchains and onion routing would require a revolution in user interface design to have a chance; otherwise it will be a playground for the technology elite

AOL used to be internet in the US, there was Minitel in France and i-mode in Japan. All of them were widely successful in their day but turned into obscurity when disrupted by more open and distributed solutions.

There’s a law which says something like “any institution always strives to become more powerful over time”. I can’t find the source of this law, but it applies to any organization – commercial or non-commercial. In their strive to become more dominant in a market or sector, they typically strive to centralization of power, influence, usage and product.

For companies this typically means they want to own part of the space they’re in and keep expanding that space. After AOL and their local competitors, the internet became more decentralized by lowering the barrier of entry for everyone. Now we’re back at a centralized internet where companies have a Facebook page with a gmail address instead of a website .

The problem with centralization is that it stifles innovation. It is an unintentional side-effect of the normal course of business, but the AOL platform and now the Facebook platform are not very fast-moving innovative places. Facebook’s mission is to sell ads and everything they do is focused on that single goal. For users this means that they want to get you to visit them as long and frequently as possible to show you ads. Don’t let Zuckerberg fool you when he talks about connecting people and creating communities. That’s just marketing, as a business they’ve just that one goal. It’s also understandable that Facebook wants to extend and protect their platform. As a company they do exactly what they should doing and their investors reward them for it handsomely.

If we step back though, only a small number of entities command 80% of your time on the internet today and this is the problem. At the same, I would argue this is the normal cycle of maturing platforms. In other contexts it’s often referred to as bundling and unbundling.

I’m a technology optimist and I’m confident this problem is going to be solved over time.Today, significant money is invested into blockchain and this could be a solution, but only if they can get it to work right. We’ve seen this before with peer-to-peer technology and that never really took off (besides downloading movies and music). Two main commercial examples of peer-to-peer technology were Spotify and Skype, but they both moved away to centralized systems – mostly because their implementations were in essence centralized, just their protocols were decentralized peer-to-peer. In those scenarios, it does not make a lot of sense. Now if Spotify would’ve been the repository of music and would allow anyone to make a client to access their database, it could’ve worked. The argument can be made for Skype. But the value in their perspective is owning the customer, the experience and the data.

The view on data and customer ownership is ultimately what needs to be disrupted for the internet to change. If companies can accept they don’t own your data or you (as a customer) and build successful business models around that concept, we can see a big shift from the centralized internet today to a more open decentralized internet.

Back to the future – hello WordPress

All of my twelve regular readers (hi mom! 👋) might have noticed that I’ve moved my blog again.

After a journey of a few years to Medium and then Squarespace, my blog is back on a self-hosted WordPress site.

I’ve been hosting my own website since 1995, but didn’t have my own domain until 2001. In 2001, I registered I remember that before 2001 you were not allowed to register a .nl internet domain as a person. Typical Dutch institutional thinking; why would a person need a internet domain? This can only lead to problems, let’s just prohibit it. It’s better that way. Of course some smart Dutch person founded a company called Ladot and offered to register the domains for you and then rent them back to you. A couple of years later SIDN (institute who manages .nl domains) came to their senses and changed their policies.

In 2001, my site run on Typo3 CMS on a Linux PC in my hallway closet – no joke.

I did this for 10 years. I even moved everything from The Netherlands to New York. It was offline for 5 weeks while my server was in a container traveling on a boat from Rotterdam to New York, but that’s ok right?

Somewhere in between I migrated from Typo3 to WordPress. In the early days WordPress was pretty ghetto. It got hacked all the time – even when you (tried to) kept up with all the updates. My site got defaced at least once a month.

In 2012, I moved to a shared host, but after years of neglect it became a administration nightmare. Upgrading was a problem because it broke my templates and not upgrading meant that my site got hacked. It happened a few times.

Sometime in 2015, Medium became a thing and it offered quite a nice place to bring your blog to life. One of their more interesting features was the ability to create a magazine. You could intersperse your own posts with posts from others and create an interesting reading experience. But when Medium grew, it became like all other social media platforms. They turned it into a feed like Twitter and Facebook. The magazine feature got scrapped. But the feed also creates problems. The biggest problem with feeds is that the harder you work, the more exposure you got but everybody else was left with nothing. I got less readership of my posts on Medium than when I hosted it myself. I’m sure it was my lack of active promotion and most probably my stuff wasn’t that interesting anyway, but it bugged me. It felt like my stuff fell into the black hole of social media.

One of the interesting later changes to Medium was the ability to redirect your domain to your Medium host blog. It worked ok and it gave me my own place. I was no longer shackled to the “feed”. But that platform also had major limitations. It was only about articles and posting. There were only very limited ways to customize your blog. That’s fine within the context of Medium, but it wasn’t really for me.

After Medium I moved to Squarespace for a bit. Squarespace is very polished and easy to use. I would recommend to anyone, but it’s also limiting and it was those limits which annoyed me. For instance you cannot experiment on Squarespace. You cannot copy your site or play with other themes, all changes can only happen on your production site. You cannot modify templates except when you pay extra dollars per month. The functionality they offer is powerful and mature, but when they do not offer what you need, you’re stuck.

I like the tinkering and missed it.

So I decided to take the plunge and go back to hosting my own WordPress.

Trying something new – RSS to email

There are quite a few blogs I like to follow and I visited them regularly to see if there was an update. But since the update intervals for each of these were very different, regularly visiting wasn’t really working for me. Some of them post once a day, some of them once a week and others only once per month or so.

A while ago I was thinking about this problem. I know there’s RSS but using a RSS reader leads to the same problem with a different interface. I rather be notified then pull since these are low volume blogs with some having a highly irregular posting interval.

Another option is Twitter, but the signal-to-noise of Twitter is in general too low for me to bother with it.

Then I came across this post from Fred Wilson where you can subscribe to his RSS feed via email. I was wondering if there was a generic solution, just enter the URL and your email to get notifications of new posts in your email. Since I won the newsletter battle I’ve a relative well managed inbox and this seems to me the perfect solution to my problem.

I found Blogtrottr which does exactly that. You enter the URL and it automatically finds the feeds (either atom or RSS). You enter your email, confirm your subscription and you’re done.

I’ve been using it for a month now and it’s perfect.  

11% of US households have cut the cable

I cut the cable 5 years and have not looked back since, but I am still in the minority. Only 11% of the households in the US have cut the cable and went exclusively to streaming. Part of the reason is the unavailability of sports but ESPN is launching a new streaming service this month – finally. It’s even reasonably priced at $4.99 / month.

I’m surprised that the number of TV viewing hours on streaming is still so low – it’s actually the same as number of households which cut the cord. You could conclude that cablecutters today are not watching as many TV hours as everyone else. I do expected it to be higher. Perhaps there’s a difference between TV watching and TV as background noise which is skewing the statistics. Also I wonder how well viewing habits of services like YouTube and Twitch are captured in these.

Mere Potential Can Be More Important Than Achievement

Zakary Tormala: Mere Potential Can Be More Important Than Achievement

In a series of experiments set in different contexts, we found that high potential can be more appealing than equally high achievement. Our studies uncovered this in situations ranging from basketball player evaluations to hiring decisions, to salary offers, to grad school admissions recommendations. We also saw it in perceptions of artistic talent and in people’s intentions to try a restaurant. In general, potential seems to engender greater interest than achievement.

This is very counter-intuitive on a rational level.

Self-Destructive Behavior of Entrepreneurs

The article Why do entrepreneurs engage in self-sabotage? tells something about human nature which in general applies to everyone, but gets super-exposed when starting a company.

In his 1985 book, Innovation and Entrepreneurship, Drucker relates the story of Alfred Einhorn, who invented Novocain, which then became popular with dentists as a local anesthetic. Einhorn held a contempt for dentistry, since it was such a small market. He felt that Novocain should be used by surgeons for all forms of surgery. General surgery was more prestigious than dentistry, and so Einhorn waged a campaign against the use of Novocain by dentists. In the end, his innovation was successful despite him, rather than because of him. According to Drucker, this pattern, where a product is undercut by the entrepreneur who created it, is extremely common.

In his book, Lucky or Smart, Peabody says it is important to be smart enough to know when you are getting lucky. And then, you have to be willing to accept that luck. This takes humility. What’s needed in an entrepreneur is emotional resilience, the kind of strength that allows a person to show grace when their ideas have been proven wrong. One has to adapt to each surprise.