This might be a slight oversimplification, but it strongly suggests that the concerns raised in my last post are not far off the mark.
First of a few posts with my own thoughts arising from the recent New Scientist ‘Instant Expert’ event.
Games and Game Theory appear to be the ruling paradigm for the current AI top dogs. Both Irina Higgins and Simon Lucas made clear cases for the choice of gaming environments as AI training grounds, and referenced Game Theory, too.
Don’t worry, I’m not going to try to argue with them – but I do think it is worth examining the assumptions that underlie gaming approaches and Game Theory, and considering these as they relate to the problem spaces which we dearly wish that AI could help us with. As you might guess, I am not sanguine… Continue reading
Pentti Haikonen is adjunct professor in the philosophy department at the University of Illinois at Springfield. An electronics engineer by training, he has constructed the experimental robot XCR-1, designed to exhibit conscious-like behaviour, and has written several books on approaches to developing conscious robots.
Kerstin Dautenhahn is Professor of artificial intelligence at the University of Hertfordshire. Her research interests include social learning, human-robot interaction and social robotics. She is the co-creator of KASPAR, a robot designed to help children with autism develop communication skills.
Making the third presentation at this event, Kerstin explained her background in biological cybernetics, and the ways that her work revolves around the interactions between humans and robots/AI, concerned particularly with systems that aim to help people.
She was concerned to be immediately clear: robots are not people.
Elaborating, she pointed out that each robot you encounter is a de novo creation, not only lacking a common biological heritage with humans – making them unavoidably alien, but without any necessary shared characteristics (either deep or apparent) with any other robot.
Further, now and for the foreseeable future (in her opinion), robots have no psychology – there is no ‘mind’ in there.
The term robot, then, is a moving target, without a firm definition (I was surprised that we weren’t reminded of the etymological origin of the word ‘robot’ in the Czech word for ‘slave’), so that any conversation about robots must be particularly clear as to terms. This, however, is difficult, because of two very strong human traits;
Irina Higgins is a senior research scientist at DeepMind, and has a background in neuroscience.
The second presentation at this event largely focused on telling a story about DeepMind’s development of AlphaGo – using this as a vehicle to explain DeepMind’s approach and give insights into its culture.
She told us that DeepMind now has 300 scientists, and was keen to emphasise the high-minded aspirations of the organisation – from its mission statement;
to its ‘intentionally designed culture’, which aims to mesh the best aspects of industry and academia; the intense focus and resources of the former with the curiosity driven open-ended approach of the latter.
DeepMind’s operating definition of general intelligence is apparently; Continue reading
Simon Lucas is Professor of computer science at the University of Essex. His research is focused on the application of artificial intelligence and machine learning to games, evolutionary computation and pattern recognition.
This was the foundation-laying talk of this event, and it was excellent – a rapid-fire but followable overview of the history and principal themes of AI research and development, and more detail on the approach currently producing the results that have been making headlines – neural networks. There was nothing here that some general reading wouldn’t get you, but it was engagingly and thoroughly presented at speed.
This New Scientist event was aimed at a general interest audience, rather than an expert one, but assumed a relatively high level of general understanding – the presentations were light on technicalities, but not shy of discussing complex ideas. I had booked without looking into the speaker’s details, trusting to New Scientist to deliver, and my trust was over-rewarded, as the presentations provided a wider range of views than I could have imagined.
These notes are provided mostly because a number of people I’ve spoken to since weren’t at the event, but were wishing they had been – they will be a poor substitute for having been there, but will hope to convey the key points and provide some links. I’ve split the event up into several posts – skim the headings and dip in to the parts that interest you – there is no grand overarching story here, folks!
Equity distribution in early-stage startups is a slightly odd subject. Obviously at this point the startup is worth nothing – or less-than-nothing, if expenses are being recorded as debts on the future company – and who wants to argue about percentage points of nothing? Sometimes the whole subject is just ignored.
On the other hand, whatever the addressable market size of the idea at hand, the spectre of founders squabbling over enormous wealth is lurking somewhere in the subconscious of everyone involved, so it is equally possible to go the other way, and invoke complex calculation methods of one kind or another, however irrationally over-fussy.
While complex approaches are arguably better than failing to address the issue at all, a simpler method is more typically adopted: if there are two founders at the beginning, they are usually assumed to have 50% each, if three, 33 1/3%, etc – as in this Seedcamp agreement template.
If they add additional co-founders, there is a re-distribution by agreement, such that the original co-founders see their percentage ownership reduced, to ‘make room’ for the new partner. The process is repeated each time a new equity-holder is added (ignoring such things as special share types – usually considered as over-complicated at early stages).
I consider that there are several problems with this:
I’ve been going to quite a few events recently which broadly come under the heading of futurism – indeed many of them have been through a reliably high quality meetup group actually called London Futurists.
These meetings deal with more-or-less mind-boggling speculations and predictions of things like robots taking all the jobs, artificial intelligences surpassing human capacities, people hacking their own or their children’s biology through genetic or prosthetic modifications, and similar subjects. Sci-fi stuff, you might think …
We’re building a medical app. Of course, Therapy-Smarter isn’t collecting deeply intimate data – just basic contact information, some physiotherapist’s notes, exercise prescriptions and exercise performance data – but nevertheless, medical data is medical data- it’s inherently sensitive, and any company that cares about its reputation needs to take data privacy – and thus data security – very seriously indeed.
So, we’ve been thinking about it fairly hard – but not in a technical way; it’s a specialist domain and we assume that we will need to pay people who know what they are doing to advise us on best practice and then get them to assess our implementation.
No, we’ve been thinking hard about security in terms of business culture, because it seems painfully clear that this is where security weaknesses really come from. That’s right – I’m saying that security weaknesses have much more to do with business culture than they have to do with engineering.