Tag Archives: artificial intelligence

🤖 Strawberry picking robots, winning users and societal challenges.

This week returns to strawberry picking robots, overcoming inertia to win users and the challenges of user generated content for children and wider society. A rather wide ranging newsletter this week then.

Picking strawberries.

When we held our entrepreneurship event for the London Symposium (an event which celebrates UK business) a few years ago, the focus was very much on robotics and AI.

One company showcasing their wares was the Shadow Robot Company which was looking to build robots that could pick strawberries. Not as easy a task as you might imagine and requires both clever hardware and software.

Something like this is tough because it is not something you can just tell a robot to do. There are too many variations. Instead a robot needs to be taught the fundamentals and then work out the best approach itself.

Embodied Intelligence, which has just come out of stealth aims to take this very approach. It does so by mimicking a human and learning from each interaction.

Eventually it will have done something enough times to figure out the best approach itself. A solid approach albeit one that is still much much slower than teaching even a baby. It seems to require a clean environment as well right now so I’m not entirely sure it is up to the task of picking strawberries yet. One step at a time though..

More on Embodied Intelligence’s approach here.

Overcoming inertia to succeed.

Inertia. Never underestimate it. If you want someone to switch from one thing to another you need to overcome it and something that you use everyday requires something big to succeed.

When Hotmail launched webmail in the mid 90s, most people were stuck with their ISP email accounts. Awful to use and difficult to access anywhere there wasn’t much inertia.

So with a big advantage and excellent marketing Hotmail became extremely popular despite heavy competition from copycats. Its viral marketing was so superior that it is still used today as an example of how viral marketing can succeed.

It was not till 2004, that someone figured out a way to overcome that inertia again. Gmail announced to much disbelief that it would give away 1 GB of space to everyone who uses its platform. It also dropped the idea of folders and introduced search. Enough to overcome inertia and woo a lot people away.

Fast forward to today though and there doesn’t seem to be much reason to switch email provider (at least so far). AOL launched Alto back in 2010, which was a nice clean email client, easy to use and with clever features. In some ways it was similar to Google’s Inbox client, which takes a similar approach of group together types of messaging and aims to streamline the email experience. Some love it, some hate it. Regardless it clearly was not enough to persuade people to switch and as a result it is shutting down on December 10th.

The next generation email client is not dead though and I suspect we are on the cusp of a new type of email client. One that integrates with smart assistants and will make email easier to deal with. One day..

What are we doing..

Happy group of kids playing at the park

Bringing up children is wrought with challenges. The world is never the same as when we were brought up so new decisions have to be made. Is the same approach as when we were younger the right approach or should a different path be taken. If a different path is taken what doe that do to other decisions? Challenging indeed.

One (obvious) huge area of change is the amount of technology available to children today. Whilst screentime is always a concern, I seem to remember being hounded away from the TV when I was young so actually this is not such a new problem.

It is actually what is on those screens that is creating the biggest challenge.

The ability to find almost any content you need on YouTube has meant that children get introduced to YouTube very early on. This is as a direct result of YouTube allowing anyone to upload content.

By removing the friction of publishers and TV channels from the mix, any creator can quickly meet the desires of an individual.

The downside it that this can be abused and technology which understands content is not as good as the technology which creates content.

So today, when it comes to apps like YouTube Kids, which visibly sounds like a YouTube safe for kids, is not really. It is not a human curated channel and means that sometimes the technology filters get things wrong.

Is that OK? At what age does it become OK? It all depends on the content, which in turn means that initially supervision and eventually teaching children to be both self-aware and understand core concepts is even more critical. This is nothing new though.

Eventually, technology filtering will catch up and then we have bigger questions to answer. This is just one more strand of a bigger problem where content appears that we (as a society) do not want it to. The obvious example today is fake news – but already this has become more broadly defined as content someone does not agree with. At some point we are going to have to define what should and should not be filtered before we drift too far towards censorship.

James Bridle has written up an excellent and detailed article looking at content targeted at kids, how automation is being used and how it is all having an impact on not just kids but the wider society. Read it here.

🤖AI and retail. 🌎Tech looking outwards. 🤝Positive Social Media

This week we look at tech moving beyond its bubble and also talking to at their first G7 summit. Tech looking outwards can only be a good thing. Then a look at AI and its potential impacts in retail before ending on some positivity in social media for once.

It isn’t a tech thing.

The mammoth tech companies are joining forces to form The Coalition for the American Dream. Google, Microsoft, Amazon, Facebook, Intel, Uber, IBM, Marriott International and other top U.S. companies are listed as members with an aim to pass legislation allowing young people who were brought to the United States as children illegally to become citizens.

A noble cause indeed. A side note to this is the breadth of membership – it is good to see tech companies banding together with companies outside their normal bubble. Less tech vs the government. For more on the coalition go here.

The first step..

The major tech companies have been under fire for a while now when it comes to spreading hate material and fake news. They have not been able to keep up with even current targets for removing this content.

Their first appearance at the G7 summit has resulted in agreement to even more stringent targets to remove extremist content. Everyone was positive with the meeting but it will be interesting to see if tech firms can actually deliver quickly enough here. More on the meeting here.

Twitter meanwhile has already released new rules around hate symbols, sexual advances and violent groups. Twitter has a long way to go here though as it is still reliant on people reporting hate crime.

The targets outlined by the G7 group effectively mean only automated systems can deliver in time. This of course means we are reliant on AI censoring our content. We already know that AI can be biased so I wonder who will be monitoring the AI.

In fact Google AI lead, John Giannandrea is just as worried about it. There is a good look at his views here.

AI thoughts

Harvard Business Review did a thought experiment on AI suggesting that AI could eventually predict exactly what we need and so deliver whatever we need just in time. It states the benefit to Amazon is that people would end up shopping at Amazon more. I couldn’t disagree more with this. For most people price and brand does play a factor in decision making.

For me the end result of this thought experiment is that should AI prediction get to a level of knowledge about us where it can predict what we want before we do, the pessimist would say we have bigger problems. The optimist in me says that in that scenario the AI should be under an individual’s control not a corporate. Of course after that we then get into debating the rights of the AI. This is the problem with thought experiments..

Whilst we are on the topic though, this is similar to the movement that says voice ordering through Alexa will allow Amazon own brands to win out.

I think voice ordering integrated with a screen will transform grocery shopping online. Browsing through countless items on a store website is numbing. But the ability to say show me all the juices with orange in them etc is going to make things much faster. There are some items where price and brand do not come into this but supermarkets have known that for generations and price their product ranges accordingly. Amazon may well execute better than the competition on this, but it will not mean the end of the brand.


Meanwhile ending on a positive thought, Facebook bought tbh last week – an app that spreads positivity across its network. It was only available on iOS and in the US and let people anonymously answer nice multiple-choice questions about friends who then received the poll results as compliments.

🤖Fintech | Robo-advisors | Bitcoin | Blockchain | Insurance

This week is all about fintech thanks to a request from Jo (HT!). I look at the current state of play, the rise of the robo-advisor as well as disruption in payments, AI and insurance. Read on dear audience.

Is fintech running for the hills?

Fintech is clearly important to the UK. London is seen as a global financial centre and a place for innovation so naturally it should aim to lead here. The government clearly agreed and commissioned EY all the way back in 2014 to look at the opportunity.

A lot has changed since then though. After three years, we have seen the start of brexit and a large number of finance jobs shift to Europe.

Things are not looking bad for fintech though. Research from Fintech Global suggests London was still leading the way in Europe in terms of overall investment.

There is a view that brexit does not hurt fintech outside of issues that are hitting startups across the board (investment and hiring for the most part). Certainly the fintech companies themselves are reporting stellar growth with many suggesting revenues will double over the next 12 months.

The Financial Conduct Authority (FCA) is really the key driver behind the success of fintech in the UK. Its focus on innovation and openness, illustrated by the focus on increasing global integration especially with China, South Korea and India suggest that fintech should still thrive in London.

Work the FCA has done with Lithuania to fast track approvals has been taken advantage of by London-based Revolut, one of the most hyped business banks allowing low cost multicurrency banking.

New faces

For years most of the innovation has been on the consumer payments and peer to peer lending. The biggest fintech companies are in this space but increasingly these companies and new entrants are looking to service businesses as well. Transferwise recently announced its borderless account aimed at businesses and aiming directly at new entrant Revolut.

Newer fintech entrants have seen their focus shift elsewhere. Alongside business banking, there has also been a rise in consumer bank accounts, wealth management platforms and insurance.

Perhaps the area that has grabbed worldwide attention the most though is bitcoin and its underlying technology blockchain (which you can think of as a public ledger of all transactions).

Last week, the value of a bitcoin powered through the $5000 mark, its highest ever and up 350% since the beginning of the year coming through some significant hurdles thanks to a civil war between the groups managing the digital currency. I discussed this in more detail here.

Since that peak, the price has fallen 20% going below $4000 in a week. This seems to be due to the Chinese government suggesting it is going to close down the exchanges in China, a huge market for bitcoin trading.

Separate to this China has also banned Initial Coin Offerings (ICOs). These are the latest vehicle for companies to raise money and they have been unregulated. You can read my previous commentary here as well as an excellent look by venture capitalist, Fred Wilson on the differing approaches of US and China and what it means.

Robo fintech

It should not be a surprise to see artificial intelligence make an appearance inside fintech. With so much data available from which to make decisions about things like spending habits and investment, the challenge has been working with the right data and in a timely fashion.

The largest players here are the investment companies (NutmegScalable Capital etc.) who have now grown large enough to be invested in by some of their larger and more traditional incumbents.

When it comes to your own everyday finances, Plum and Squirrelwere good mobile versions of traditional savings approaches, whilst Moneybox was a more innovative way of making you save without realising by rounding up the cost of items and squirreling away the difference. Nothing very intelligent though.

Last week, US based Pefin announced it is using artificial intelligence to help people make better financial decisions when it launches later in the year. The challenge is it needs three months worth of spending data to get started though we are not far away from EU regulations forcing the banks to open up access to this data (the UK is also going to adopt these regardless of Brexit).

Insurance with a conscience

One of the most profitable and complex industries is insurance but there is plenty of innovation happening. The most well known of these is Lemonade, which is focusing on reducing premiums, donating to charity and increasing transparency. Oh and it also uses artificial intelligence.

It makes money by keeping 20% of all premiums and sets aside 40% for insuring itself against major claims with the remaining 40% covering claims.

Unlike traditional insurance companies, any money not used for claims then goes to a charity of a customer’s choice. It has captured the imagination of consumers though still small and only available in a few states in the USA. After the first year though, it delivered $53,174 or 10.2 percent of first year revenues.

Where does it use AI? In chatbots, using them to automate customer interactions, though not many actually complete without human intervention. It is a mobile first company so its primary interface is through its mobile app though it does have a website.

If it succeeds in getting the chatbots to take over a large percentage of customer interaction, the company would be attractive to existing insurance companies regardless of the success of its business model. Of course, the insurance companies may want to try and buy the company and shut it down if it becomes too successful and undermines their very profitable current model.

The model is inspiring other insurance startups to adopt a similar approach. A new yet to launch pet insurance company in the UK is looking to follow a similar model and I would be very surprised if there weren’t others.