Dogs are a fun source of metaphors for machine learning. We can see the learning process much more clearly than we can in ourselves, but we can also sometimes see it going wrong in odd ways - we don’t *actually* know what the dog has learnt, we just think we do. And we can’t ask
On the other hand, dogs do have general intelligence. But it’s not the same as ours, and again, sometimes you can see a dog failing to work something out that seems really obvious. “Why can’t he work out he can go through the side door if the back door is closed?’
John Smith @JSmithpolitics1
How is Tencent & Ali Baba's machine learning compared to Google or other top American firms? Are they making any advances beyond what we have in the US? Are they comparable or behind? Thanks
Sometimes I think ‘personal data’ is almost as unhelpful a term as ‘artificial intelligence’. It implies that all data is the same, all has meaning, and can all be used in the same way.
This gets you to statements like ‘you should own your data’, which sounds impossible to disagree with until you ask what it means. ‘I took that bus’ is ‘my’ data, but are you telling a transit authority it can’t record which routes are more popular? Whose data is that?
‘I should be able to take my data anywhere’. Agreed. But if I export my LinkedIn profile and take it to Snap, what happens?