Global Technology Trends that will influence Africa in the Next Decade; Media, Bitcoin, MicroServices, Machine Learning (AI), Augmented Reality & Virtual Reality.
6. justin spratt, 2016justin spratt, 2016
1 billion people across 54
countries
Mobile penetration 90% - half
smartphone by end 2017
Natural resources make up
~40% of African GDP
Only 24% of Sub-Saharan
Africa has electricity
More than 40% of Africa’s
population is under 15
15% of the world’s
population, yet consumes
only ~4% of Global GDP
E-commerce by cash-on-
delivery or click-and-collect
due to underbanked
7. justin spratt, 2016justin spratt, 2016
African Macro Development Framework
5. Media - “religion media opiate of the masses” - karl marx
Mokoko Slum, Lagos - 300k people, ~satellite dishes
4. Education
“By 2021 half of there world's learners will be in Africa” - Steve Haggard
3. Transport
Unlocking the Economic Potential of the continent
2. Health
You cannot teach children if they are sick of hungry
1. Infrastructure
$100bn needed pa/$50bn currently - Economist
Burkina Faso, 25% of average salary required to charge phone - NGO, 2014
9. justin spratt, 2016
My
Prognostication
Principles
1. We over estimate
technology in the short term
and underestimate it in the
long term.
2. Progress happens at the
edges. Often it looks like a
toy.
And, don’t be too quick to
judge. The first
pornographic novel was
written over 150 years
before the first scientific
journal.
14. justin spratt, 2016justin spratt, 2016
“40% of today's
companies on the S&P
500 will no longer exist in
10 years.”
A study from the John M. Olin School of Business at Washington University.
21. justin spratt, 2016justin spratt, 2016
VR
- accelerate learning through real life
simulations
- computer gaming
- ringside seats at sports events
- combined with robotics, the
healthcare benefits could be huge
- sales engagements
- travel the world - google earth
comes to life
- Journalism - BIG empathy
AR
- education, accelerated learning -
rural youth
- compress and accelerate the
learning curve
- market data for farmers in Africa:
yield, pricing when trading
- games that get people off the
couch (Pokemon) and exercise
- data at your fingertips in intelligent
ways
30. justin spratt, 2016justin spratt, 2016
"It is believed
there are more
possible game
variations than
atoms in the
universe"
31. justin spratt, 2016
https://biotechin.asia/2016/09/03/machine-learning-help-revolutionize-cancer-
diagnosis/
Conventional tests failed to
show any sign of the
cancer. The doctors turned
to IBM’s Watson (AI).
It cross referenced this
case against 20 million
oncological records, and
concluded in a mere 10
minutes that the patient
suffered from a rare type of
leukaemia…
they started a new
treatment regimen, and the
lady survived.
The very best
physicians are 10x and
even 100x better than
the average.
Yet every physician is
limited by the number
of cases they’ve seen
in their lifetime…
What if we could make
the very best
extensible using AI?
36. justin spratt, 2016
Immutable record of account
decentralised (Peer-to-Peer)
Open Source (OSS)
cheaper settlements and fractional
interchange
Permissionless innovation (a’la web)
Think of it as the plumbing of digital
payments on the internet
untouchable by political machinations
justin spratt, 2016
I use this visual because people don't realise that the internet came from a concerted effort by the US Defense Department to ensure their ability to execute wars in the event of a nuclear war.
The money, which came from the government, was tax dollars.
Lesson is twofold:
innovation happens at the edges
while free markets are essential to clear and allocate capital efficiently, government can play a pivotal role in funding and creating the right context for markets to thrive.
Silicon Roundabout - Old Street, London.
Recently spent 18 months here getting up close and personal with some incredible startups.
Until Brexit, the most prolific technology venture capital locale in Europe.
Arguably, the best Machine Learning universities are in England.
Anecdote:
It is a place in which a shanty town of almost a million people, lies on the bay of Lagos, off the electricity grid, with no running water, but with almost 25,000 satellite dishes.
Burkina Faso
This is how I see development in Frontier Markets like Sub-Saharan Africa.
While the development path is linear in importance, each one can have a mutual reciprocity and stimulate the others in a quasi feedback loop.
In trying to see around corners, I have learned two things (above).
Indulgences (paying money to get out of purgatory) via Gutenberg machine was the economic model the delivered the economies of scale for this important technology.
Only then came the printing of the bible en masse.
Source: Journey of Homo Innovaticus.
James Hadek.
Source: Davos discussion, WEF
We will see step change and a proliferation of new S-curves in the next generation as Software and Artificial Intelligence augment to cofound even the most prescient futurists.
Seminal article by the founder for Netscape, Andreessen, predicting the dematerialisation of balance sheets, lead by technology companies.
Netscape was the first graphical user interface (GUI) browser
Dematerialisation and subsequent deflation of prices due to technology.
The success of the past blinds business to what is need to win in the future. It hires people and builds processes with the initial model of success as a blueprint, and then that perpetuates the establishment into the future. This does not work.
Read:
+ Schumpeter, read ‘creative destruction’ in, Capitalism, Socialism and Democracy.
+ Christensen, Innovators Dilemma.
+ Marshall Goldsmith, What got you here, Wont get you there.
Software delivered by Internet technologies will disrupt every industry.
Every company will need to be a technology company to survive.
Financial and Health have huge regulatory moats, hence the largely protected monopolies that still exist. This is changing, albeit slowly.
Industrial sectors are the mixed bag. Lots of robotisation (software controlled hardware) but with very low margins and concerns about large scale job losses, large swathes have been left untouched - but their time is nigh too.
General Electric, a B2B, industrial conglomerate, is one Snapchat and is funding Native Advertising on Buzzfeed.
To date, B2B marketing was thought of as a combination between Sales, Events and targeted White Papers.
So what is going on here?
When AI meets voice. Voice is often - when it works - is almost always the best user interface.
Voice is arguably the most human of traits that separates us from other species.
Artificial Intelligence delivered by bots in IMs.
Conversational Commerce.
It is a strategy, lead by Google and Facebook, to try and reduce the Apple hardware hegemony.
Above is the developer kit.
Oculus renders the image in stereoscopic 3D through fancy lenses, tricking you into believing you’re looking at a real environment and not a screen mere inches from your eyes.
HTC Vie is the closest competitor.
Playing war games in VR have been known to cause PTSD.
There are many documented cases where people have cried, in the belief that what they are seeing is so real.
Ditto - rarest pokemon.
Pokemon sales and licences alone have resulted in Nintendo earning a whopping $57.65 billion in revenue as of 2015. This a game franchise that started as a card-collecting franchise 20 years ago.
Progress always happens at the fringes.
I think AR will be the biggest boon for Africa in the next 10 years, allow the continent to compress the learning curve and in some instances, replace it, by having data, facts and knowledge at hand.
arguably only a few companies in the world have enough data to make the big breakthroughs in ML.
google, facebook, amazon, microsoft.
apple could be, but it has perennially underinvested in its software services.
In the last 4 years there have been significant step-changes in the field of AI.
Most notably, the ImageNet competition which was the time when machines could discern objects in images better than humans.
In the AI field, the current area most progress is being made is in ML. It is the Supervised Machine Learning where all the progress is being made, through Deep Learning & Neural Nets.
Large data sets are the key success factor in Machine Learning. Computers can only really get very accurate through pouring through unimaginable amounts of data. (There are startups being developed that are trying to hack this by simulating data, but it remains to be seen if this works.)
As such, there are only a few organisations in the world that have the raw inputs to derive value from ML in a commercial sense. This is why Elon Musk founded OpenAI.
It is very telling, that data is the key ingredient, when Google open sourced its Machine Learning platform. It is remarkable that an organisation would give away such an important piece of IP. It is a tacit confirmation that huge amounts of data, for now, is everything in this field.
"It is believed there are more possible game variations than atoms in the universe”
The point here is, that unlike Chess, you cannot programme every possible move into a computer. It needs to develop it’s own “ability” to play the game.
In this case, DeepMind, was a set of ML techniques that taught itself to play and win the ancient Far Eastern game of “Go”. It taught itself to win and beat the greatest player in the world, Lee Sodol.
thesis: “every physician is limited by the number of cases they’ve seen in their lifetime.”
also, more accurate diagnosis for treatment. we kill a lot of cells by spray-and-pray
also, depression and schizophrenia
deep-learning startup Enlit: “algorithm recognizes lung cancer with 50 percent more accuracy than human radiologists.”
InvitroCue are revolutionizing cancer diagnosis
https://medtechboston.medstro.com/blog/2016/05/24/16045/
bought by twitter for $150mn
14 people, 11 engineers
18 months old, no revenue
Magic Pony specialises in creating algorithms that can understand pictures. It means we will
Insurance-as-a-service.
Price efficiency is coming to everything.
Why is cryptocurrency going momentum?
One attractive use-case is remittances:
$90bn in remittances into Africa.
$20bn intra-Africa.
Average transactions costs for remittances are more than 10%.
There are some known instances where this has been as high as 17%.
MicroServices, in simple terms, is unitising every part of computing barring the very unique set of software instructions that create value - the stuff that actually differentiates a businesses operations or go-to-market. The theory is that we can drag-and-drop pieces that have been built before, many times, as we need it.
It is also immunises the software stack. If a problem starts, it is likely to be ‘contained’ in set of services, known as a container.
For Africa, this means we will not to need to train engineers and developers in a huge amount of what used to be very difficult stuff. We can focus their skills training on the very top of the stack. Troubleshooting complex systems becomes an order of magnitude easier.
MicroServices use what is known as “containers”, which like moving house or product, provide a very neat way to compartmentalise things that should be go together.
In much the same way, in technology, you can use containers to put specific services that would go together in one container. And in software you would have a series of containers that make up the software and infrastructure stack. These container become independent of each other, in essence immunising the rest of the software should one element become a problem. In such an instance, only the container the service resides in would be become problematic and need attention.
Docker is the most popular solution for MicroService containers.
Lambda, Amazons version, is new and slightly different. It allows the physical infrastructure to bought as ‘contained’ units. So now you have infrastructure and software services isolated into groups that work together. In an anthropomorphic way, it is making complex computing akin to building with lego blocks. This will lead to a significant reallocation of resources higher up the stack where the business value resides.
This is one example of where expensive software can be unitised and delivered at scale across many businesses. Currently, most businesses build their own unique stack. This is an old-model and many digital agencies and systems integrators will increasingly compete to remove these cost and operational inefficiencies.
We will see this across sectors. No longer will bespoke software stacks be du jour. Only very top layer of software, the high-value differentiated piece, be allowed to be bespoke for businesses. This will free up huge amounts of IT and capital budgets for business.
The marginal cost of computing is tending to zero.
This is function of Moore’s law and increasingly efficient use of energy by big data centre operators. Google, Microsoft and Amazon have been investing in this infrastructure for two decades and so now there is increasingly vast amounts of computing power at scale.
As Lambda shows, there is so much infrastructure, that it makes economic sense to even sell it in CPU cycles or MIPS (millions of instructions per second). Remarkable.
This is an example of the MC of computing tending to zero.
In 2000, decoding a genome cost ~$100mn
In 2016, decoding a genome cost ~$1k
The is the result of huge amounts of parallel computing coming onboard over the last 20 years, driving huge economies of scale efficiences.
The key message here is: even when you have the best innovation for decades, unless you are able to commercialise it, you will die.
PARC (Palo Alto Research Centre) was Xerox’s innovation hub at the time. Xerox was the most profitable company in the world at the time. They invented: the mouse, computer graphics (bitmaps), the GUI, the laser printer, Windows, typeface (fonts), ethernet (modern networking) and many others.
Steve Jobs & Bill Gates visited PARC and were able to take to market the innovations, creating the foundations of the Microsoft & Apple empires.