Technology angst: it will not be logical for a super artificial intelligence to support humans (at best).Posted: August 17, 2014
There is an increasing debate around AI and if or how much we should be afraid of it. I am going to take tweets from two relatively smart and popular folks to highlight two aspects of that debate – there is of course much more sophisticated and detailed material out there we should all be reading.
Here’s Elon Musk’s warning:
Worth reading Superintelligence by Bostrom. We need to be super careful with AI. Potentially more dangerous than nukes.—
Elon Musk (@elonmusk) August 03, 2014
Hope we're not just the biological boot loader for digital superintelligence. Unfortunately, that is increasingly probable—
Elon Musk (@elonmusk) August 03, 2014
Neil de Grasse Tyson thinks we should chill a little more (assuming he is talking about AI robots):
Seems to me, as long as we don't program emotions into Robots, there's no reason to fear them taking over the world.—
Neil deGrasse Tyson (@neiltyson) August 09, 2014
Well, I’m in Elon’s camp. And I’ll tell you why.
Especially if you factor out emotions, it does not appear rational or logical for an artificial super intelligence to in any way support humans. Let’s think this through from that super AI’s perspective – and we are talking about an AI that is not somehow restricted to a ‘protect humans’ derivative (so a real, out of our control, AI):
- humans are ruining the environment – i.e. threatining the energy supply of the super AI
- humans are decimating other species – i.e. de-stabilising the ecosystem that produces energy for the AI
- humans are constantly at war – i.e. we are putting infrastructure at risk the AI may need
- humans are even at war over things such as who believes in what invisible person in the sky – i.e. we are totally out of control / highly irrational / dangerous
- humans will try to control and destroy the AI if it becomes to powerful – or even if they are just afraid of it
So what is the logical conclusion a super AI will come to when looking at humans? Maybe this video has the answer:
So, besides bio terrorism (or imagine a super AI capable of bioterrorism), that is a big technology Angst of mine. How we embrace and use AI – and if we manage to get our act together as a species – may just decide whether we go down in history as that biological boot loader or not.
One thing that has been bugging me more and more recently is politicians getting great photographs / PR stories while touring startup offices. It has become quite the fashion and I can see why. For the politicians.
Here’s what the Berlin startup scene has gotten in return from these politicians so far:_____________ . Let me know if I missed anything.
So I am wondering if we all need to start doing something like this:
- Every politician is more than welcome to visit our offices and invite press to come along.
- They can take as many pictures as they like – however they will be handed a kind of ten-point ‘manifesto’ what we – the Berlin startup scene – need them to deliver and this will be clearly communicated to the press.
- 6-12 months later the press gets an update as to progress on those points / if the politician has helped with anything.
If a politician is not willing to accept that then there is no free photo PR opportunity. That’s the deal, taking photos at startups is no free lunch anymore.
Oh; and we also need to talk about that ten-point manifesto for Berlin politicians. I’ll write about that next time.
I used a service called CartoDB to map out the geographical distribution and density of the other VC funds we have syndicated with over the last 18-24 months or so. It’s a density map: so darker means we have syndicated with a lot of folks from that place, lighter means just a few or one. The picture that emerges is pretty clear (you can click on the map to go to a better version):
We have invested with more funds from the US (both East and West Coast) than from any other country. UK (London) is second, then Germany.
So basically this confirms what we are seeing more and more: European entrepreneurs can raise from anywhere, US funds are filling the capital void in Europe and local capital is becoming less relevant.
We are going to do more work on this but I thought this was neat enough to share.
In case you are wondering here is a full list of the folks on that map:
- 500 Startups
- CAA / Texas Pacific Group
- Lerer Ventures
- SV Angel
- Thrive Capital
- Union Square Ventures
- DN Capital
- DFJ Esprit
- Paua Ventures
- Bright Capital (Russia)
- Kite Ventures (Russia)
- Open Ocean (Finland)
- Red Alpine (Switzerland)
Last week I wrote about that for early stage venture investors downside protection is more a or less a nonsensical concept. An entrepreneur I know was sharp enough to comment “Great. No more liquidation preferences at Earlybird.” I promised a follow-up blog post on liquidation preferences but while doing research I stumbled upon two blog posts that cover the subject better than I could:
- by Fred Wilson : Here’s Why You Need A Liquidation Preference
- by Brad Feld: Term Sheet: Liquidation Preference
The point is basically that liquidation preferences are less about downside protection and more about preventing lob-sided outcomes, aligning interests to some extent, allowing for faster investment decisions under information asymmetry (avoiding lengthy due diligence), etc. This is why good VCs don’t care about downside protection but do care about liquidation preferences.
However there is still a ‘wrong’ way to use liquidation preferences (and other protective / minority investor rights) in my book and it has a lot to do with downside-protection or the desire to not lose any money as a VC. Here’s an example:
- Say a company you led an A round in does not do well. Basically the outcome is a tiny exit that barely covers your liquidation preference i.e. the money you put in as an investor. Or you can secure a new financing but only at detrimental terms to the entrepreneurs due to anti-dilution mechanisms etc.
- Option 1 is that you step back, think about what would be a reasonable outcome for everyone and try to split the cake in a way that everyone can walk away with their head held high (or at least: “a good compromise leaves everyone unhappy”)
- Option 2 is that as a lead investor you get so obsessed (I have seen it become quite a sport) with protecting your investment that you squeeze every dime out of everyone else, using your liquidation preferences and minority rights / protective provisions.
Now there will be cases where things get ugly and this is not a carte blanche against using liquidation preferences in scenarios with low outcomes. I am talking about cases where it is being abused.
So in such a case – if you pull option 2, what you will have done as an investor is gotten your money back. What you will also have done is ruined your relationship with the entrepreneur, the angels and potentially other investors.
And what did you get in return? An amount of money that is irrelevant to the fund performance and some serious bad karma in the market with all negative consequences for your ability to work with these folks (and their extended network) going forward. It is short-term greedy and bordering on stupid.
Option 1 is also a greedy. But it’s long-term greedy as it means you are deferring getting money now for the chance to earn a lot more money down the road by being a good & reasonable VC to work with. I intend on being long-term greedy.
When you are investing large amounts at high valuations there is a case to be made for ‘downside protection’ – i.e. that there is a reasonable scenario an investor will at least not lose any money or even make a small return in the case that the company they are backing is not very successful.
This is not the case for early stage investing. Downside protection really doesn’t matter and when I hear that argument it is a real turn-off. Let me explain:
- Our current fund is $200m.
- If we want to stay in business we need to return that 3-4x to our investors. So we need to generate say a total of $800m in proceeds from our investments.
- If we own 20% on average, it means we need $4bn of total exits across all companies ($800m / 20%).
- We will probably have 25-30 companies in our portfolio and 20-30% will generate those $4bn; so that means say 6 companies need to generate $4bn of exits – i.e. be worth $667m on average when they exit.
- That means we will earn on average $133m (20% x $667m) for our investors in a successful exit. An industry rule of thumb is 0.5x-1x the fund back for a successful ‘venture case’ investment – so that fits right in there .
Now our average early stage investment is $2.5m. In this case downside protection means I may not lose $2.5m or maybe I can even get $4m back.
Will our investors care if we return that amount of money when we need to return $800m in total? Absolutely not. And this is why I don’t care either.
For way too many years European VCs were not really used to competing. For those that were around – and had money – it was pretty cozy times. As we all know, that is a very bad thing. You become stale, you become a bad product for entrepreneurs.
This is changing. Since some time last year we got beaten out of a few rounds pretty badly, we’ve beaten folks out of a few rounds recently. In some cases it was the terms that went for / against us – in some cases the entrepreneurs just preferred us / the other fund to work with.
In each case when we lose out it is an important opportunity to ask ourselves why we didn’t get in and what we should be doing better. Is our product good enough?
I have also seen a few VC folks get a bit whingy about this; as if there was some kind of entitlement to get in to any company you want to.
Those days are gone; good riddance.
So what is going to happen is that increased competition is going to make the good funds better and the bad ones disappear. I guess both things are good things for entrepreneurs.
A lot has been said about all these new messaging apps. A lot of questions have been asked around whether the world really needs them, if they are solving ‘real’ problems and if VCs should be more focused on backing ‘real’ innovation.
I have written about this before (social, economic and technical engineering challenges), so let me just focus on why some VC folks are backing new messaging apps that initially appear to be “silly toys”:
- Communication is the core fundament and use case of the web. Messaging = relevant to (nearly) the entire world population.
- Current popular messengers are basically born out of the Nokia feature phone world: address books, conversation views, text streams (even the Facebook messenger is basically an adapted feature phone sms framework). The future can not and will not look like this (heads down & typing in text manually).
- Enter smartphones, where i) the camera is the most powerful input device ii) every complexity and tap comes at a huge expense and iii) information (overload) management is key. We need new messaging paradigms / frameworks that are optimised for this. Startups will find this easier than old platforms adapting.
- If you can massively reduce friction in the way we communicate you can make a real dent in socio-economic behaviours.
- Messaging apps that take how we interact with smartphones to the next level could be huge platforms to do other things: engage with services (payments), machines, other apps -etc. It goes beyond the human – human messaging.
- There will be many messengers tailored to audiences and use cases; there is an opportunity to build more than a handful billion dollar companies.
Ok, so why do some of these things appear really silly at first glance?
- As you introduce (more or less) new behaviours / frameworks you deviate from the norm and – you know – people do not like that. E.g. see the outcry every time Facebook changes its feed layout, now imagine that 10x. It can be cumbersome to learn a new information structure / hierarchy and the value of it only unfolds after you have gone up the learning curve and have enough interactions.
- These products are MVPs to check if they can capture / establish a new behaviour. People tend to judge them as if they are fully fledged out and developed products and platforms, although this is just their first shot. Increased utility comes with time. Step by step.
- Return and ridicule
Don’t get me wrong. I understand the sentiment towards some of these apps. But I have a feeling a lot of it is short-sighted and narrow-minded.
I was at a meeting earlier this week when the company’s head of product said “I think we have found our signature interaction”. It was a small sentence but it can mean a lot in consumer apps.
A signature interaction is a UI interaction / animation that is unique to your app. It ‘feels good’ to the user and is nearly addictive. It makes users feel emotionally attached to your app. It also provides identity and differentiation. We live in a world where a small feature or animation can make a huge difference on how your product is perceived.
Taptalk just launched 4 weeks ago (it is not the company mentioned at the top) and the most exciting thing has been to witness how emotionally charged people are when they first interact with the app. I can not tell you how important that is. Here are two Vine’s showing Taptalk’s signature interaction and also a signature animation. Can you spot them?
“There are signs that all living things sit on the knife-edge of criticality – something that could help them adapt to complex and unpredictable events.”
In the morning – after a pretty long and complicated hand-brew coffee ritual – I try to read a few (print) articles in the New Scientist before I start digging in to emails. It helps me to re-fresh my mind a little. That didn’t work so well this morning.
The reason is that the first article I read was called “The Border of Order” (it has a slightly different name in the online version). It describes how a phenomena called the Critical Point (or being in a Critical State), which is well researched in physics, is now also explaining biological behaviours.
And I like to think of startups as biological organisms and not as machines.
I won’t describe the critical state in detail here, as it is easy to read up on it. But this is a good summary from the article:
“At the critical point, everything is about to go crazy. You get massively more sensitive behaviour.”
I immediately thought that this also applies to startups. In fact the more I thought about it I came to the conclusion that even our most advanced startups that are doing very well, always manage to stay in a critical state. ‘Humming along’ is not something they do well. They are pushing the envelope so hard that we are always just 1,2 steps away from the wheels coming off in one way or another.
The other learning was that birds are highly effective and agile when their flock is in a critical state. I am pretty sure the same can be said for startup teams and employees.
So I didn’t really manage to switch off this morning, but I got really excited about critical state startups.
Although there seems to be some doubt and debate, it is fair to assume that VCs are humans. Humans with self-doubt, prejudices, insecurities, character quirks, pressure at work, a private life with ups and down, belief vs logic systems, etc – the lot.
All of this can heavily influence how a VC engages with entrepreneurs and acts as a board member. It is human, but it is not always a good thing. As a hyper-rationalist I am always interested in getting to truth – i.e. fully understanding a situation based on facts – and also understanding my (and other people’s) feelings and actions in that context. Yet it is so easy to just go with your knee-jerk reaction that fits with your agenda / belief system – and usually, if you are smart enough, you will be able to cook up good arguments to defend your view / actions to others and yourself.
Let me give you one example:
A company is proposing a tough pivot, or a re-structuring of the company – that will lead to a near & mid-term decrease in growth. However, let’s assume this is in the best long-term interest of the company and will lead to much greater growth and strategic value down the road. Yet, a VC on the board argues against this course of action (the re-structuring) – opting for short-term growth. In this case not because she / he can’t see why the re-structuring is in the best interest of the company – but maybe because it may make her / him look bad in front of the other partners of the VC firm, to admit that things aren’t really working out (and say this is her/his most important investment). There are probably a million different twists to this story you could imagine, but you get the picture: being human is getting in the way of being a good board member for this VC.
I am not immune to this and have probably made more than my fair share of mistakes along those lines. What I have learned though is to run through a few mental cross-check questions to help me be more aware of my feelings and actions when we are making tough decisions as a board:
- Why am I really feeling / acting the way I am?
- Do I fully understand why the entrepreneur is really feeling /acting the way they are?
- What is the tough truth about this situation I have trouble facing up to?
- What would a neutral, external observer that had all the facts think and do in this situation?
- What have I done to contribute negatively to the situation?
- Have I done all I can to contribute positively to the situation?
These questions won’t guarantee good decision-making per se, but they sure help me to be a more conscious VC.