Your Artificial Intelligence Low-down
Welcome to the next blog post in a series alongside NatWest. This month we will be exploring everything to do with Artificial Intelligence: providing insight into the most recent AI news and events of note.
Why Invest in AI, by Nare Vardanyan — Entrepreneur in Residence at AI Seed
Deconstructing misconceptions about investing in companies specialising in Artificial Intelligence.
Humans are big fans of overhyping phenomena, then going through periods of rationalisation, hence disenchantment.
Overhyping leads to huge valuations of certain assets classes that are at the dawn of revealing real potential, followed by disappointed withdrawal of all resources and change of focus.
Artificial Intelligence, a term coined in the midst of the last century, but one that has been in the imagination and folk tales of our species for thousands of years, has recently become a massive target for hype, expectations and investments.
The race for Artificial Intelligence is on, with the biggest companies (and the best performing stocks for that matter) in the world betting massively in this field, swooping talent, infrastructure and data. This is a known fact, a cause for huge societal worry, considering the lack of conceptual and ethical frameworks to control, measure and even fully understand the effects of A.I.
Today, as we are on the down curve of the Gartner hype cycle for A.I., it is the best time to realistically look at the value creation opportunity of A.I. first businesses and deconstruct the logic of investing in them early.
First of all, let us define what an ‘A.I. first’ business is:
Companies who are using cutting-edge Artificial Intelligence techniques to address complex problems, that were impossible to solve before.
At AI Seed we invest in ‘founders creating superpowers for the software that is eating the world’. While it seems like a no brainer at a first glance, it can be more contrarian than you think, when we dive into the depths of this.
Here are 5 common arguments that are brought up when insisting this is a bad idea and some of our thoughts around them.
One might argue that just like any other technology, A.I. is going to become ubiquitous and there will be no A.I. startups to invest in.
Yes and no. This is true about any cutting edge or emerging technology, but the timelines are usually underestimated.
Several investors have claimed both publicly and privately “A.I.” is not really a category but the next phase in an evolving enterprise software stack. We believe the coming decade of automation is more akin to the technology waves of the web, mobile and some would even say electrification.
The barriers to entry are high, A.I. talent is still research driven and scarce, despite the openly available open source tools and frameworks and the wave of democratisation of techniques and methods. Hence the opportunity for creation of proprietary advantage at the earliest stage cannot be overlooked.
Any large enterprise can become ‘A.I. first’ hiring researchers and engineers. Yes, in theory, not so easy in practice.
Google can. And they are doing that. Perhaps, Amazon, Facebook and Microsoft as well.
However, for your average large enterprise, this requires a massive mind shift, being able to understand the types of problems A.I. can solve and creating completely new products and services. As long as the chasm is there, there will be opportunity for A.I. startups to harness the power of early proprietary advantage.
Data is king. Small businesses with no data have no chance. Big tech businesses owning all the data are natural winners.
A fair, but flawed, statement if we decompose it.
Data networks effects are powerful, but data itself is not a scarce resource if you think about it. These are spheres that have not had massive net productivity gains through the last decades. Our founders love them. We love them too.
Privacy and open data is also becoming more important, as citizens get more and more concerned about their data, where it goes and how it is used. This is one the trends supporting the excitement around decentralisation. While complete decentralisation does not exist, there are ways to improve on it. That seems to be where we all are headed.
Open AI, Ocean Protocol, OpenMined are among the early evangelists of this approach. We are believers too.
Does this mean our startups will lose proprietary advantage? Not really. It means they will be in an open competitive environment to fuel their solutions with data and build better ones to access market opportunity and grow.
A.I. first businesses like any other business, need to gain commercial traction and validation. Hence, the metrics to value them and invest, should be no different.
True, but… It takes more time to build an A.I. product ready for production, than for an average minimum viable product. Many A.I. startups will die before they get to this stage due to lack of capital.
Hence, AI Seed goes in early, supporting exceptional founders with a unique understanding of the problem they are solving, allowing them to focus on building, instead of worrying about bringing early sales for a ‘non-existing or poor’ stack in order to get investment.
Early exits are the destiny of any A.I. company. They are normally not massive either, which prevents bigger funds considering A.I. as a relevant category for investment.
Well, this point is definitely the hardest to argue with, as it is heavily supported by data. However, it is somewhat contradictory to the initial truth about any company becoming A.I. company sooner or later.
We strongly believe, it is too early to see A.I. first businesses becoming huge success stories capturing significant market share. As we have not seen that happening, we prefer to think it will never do.
First of all because the larger public and enterprise are just realising the power of A.I. to transform everything, we are stepping into the area of building, instead of purely evangelising just now. Also, if there is not enough later stage capital to power the growth of these businesses allowing them to move to category-creating or dominating phase, we will never see it happening.
Thirdly, this is a great case for a new wave of full-stack businesses with optimised system that makes them capable of out-competing everyone else.
Finally, the scientist-entrepreneur is far more rare than either an incredible scientist or a brilliant entrepreneur.
Yes, yes and a million times yes. That is why we back them from day one.
Our team started around the premise of knowing what we do not know. We back people who know more than we do and have curiosity and drive for learning, like we do.
We feel privileged to back founders who are risking it all in order to create their version of the future.
To crash Nare’s points, have an informed debate, get introduced to our team or portfolio at AI Seed, please do get in touch.
This month’s AI Article Update
Do take a look at the amazing offers for AI start-ups contained within the report.
The map that states the UK is the 2nd best place in the world for AI talent…
CB Insights take a look at 13 artificial intelligence trends reshaping industries and economies.
- ‘The State of Artificial Intelligence’, by Andrew Ng
Professor Andrew Ng is the former chief scientist at Baidu, where he led the company’s Artificial Intelligence Group.
- The Founder’s Guide to Raising a Series A Venture Financing, by Justin Kan — CEO of Atrium LTS
The article provides great advice for Scale-ups raising a big VC round
- What does it take to raise capital, in SaaS, in 2018?, by Christoph Janz — Managing Partner at Point Nine Capital
Another article about what it takes to raise a Series A in SaaS.
- Want to be a founder? Go get a job, by Joe Kraus of Google Ventures
AI Startups to watch…
re:infer raise a $3.5m Seed round
re:infer announced on 21st February that they have raised $3.5 million in Seed investment, led by Touchstone Innovations with participation from Crane Ventures, our own fund AI Seed Fund, Dr. Jason Kingdon, and Seedcamp.
re:infer is bridging the gap between human communications and IT infrastructure in the enterprise — a perceptual system that listens to multi-channel unstructured communications data, discovers inefficient processes, provides customer intelligence and delivers automation.
Factmata raise funding to expand in the fight against fake news
Earlier in February, London-based Factmata closed a $1m (£720,000) Seed funding round, led by tech entrepreneur Mark Cuban, as well as high-profile investors such as Twitter co-founder Biz Stone, internet entrepreneur Sunil Paul and Craigslist founder Craig Newmark.
Factmata will use the funding to expand in their quest to tackle the proliferation of fake news and extremist content online.
Maxatta
Maxatta are using Robotic Process Automation (RPA) for investment banking front-office. They are on our fund AI Seed, and we are looking forward to working closer with them.
Upcoming AI Events
PAPIs Europe — 4–6th April 2018
After previous events on 4 different continents, PAPIs is returning to Europe this Spring! Capital Enterprise are proud to be a co-organiser of the conference.
It’s a great opportunity to connect with international ML experts who will speak about real-world ML applications, techniques and tools. You’ll hear from practitioners at Google, Amazon, Scikit-learn, Adobe, Element AI, BigML, Dataiku, Rapidminer, and more selected via a rigorous blind-review process.
PAPIs also features the world’s 1st AI startup competition where the jury is an AI! With the chance to win a £100K investment from our very own AI Seed Fund)…
Join us on 4–6 April at Level 39 and UCL!
Machine Learning Conferences in 2018
A great articles by Alex Kistenev, CEO of Standuply, listing 253 business and academic events this year all about Artificial Intelligence and Machine Learning.
Special thanks to Nare Vardanyan, John Spindler and Gracie Jones
