Alan Kay wants us to know that the world is not always as it seems. We start with the classic example of a long table and a wide one with the same area – even showing the transformation of one area into another, it’s a deeply counterintuitive truth. We’re easily fooled, he reminds us, and we like to be fooled, which is why we go to the theatre, watch magic shows and such.
Quoting the Talmud – “We see things not as they are, but as we are” – reality is a form of waking dream, where our way of viewing the world can be a barrier and a distortion. The complex might be made simple, the simple might be quite complex – we are a noisy channel.
To enhance our abilities within this noisy channel, we build “brainlets” – tools to let us correct the noisy channel. They include tools like telescopes and microscopes, as well as reasoning tools like logic and math.
Kay wonders whether the models we build are too simple or overly simple. He looks at Hans Rosling’s Gapminder and concludes that the models are quite accurate. But he critiques the David Bolinsky film for oversimplifying the actual interactions between molecules – when molecules match in the film, it looks inevitable. But in reality, those molecules are spinning millions of revolutions a second, and its their revolution that allows molecules to link with each other.
Kay claims that we can confuse adult sophistication with actual understanding of a mathematical or scientific principle. He shows us footage from a magnet school in Los Angeles, where children play with shapes, building larger versions of a shape from small shapes. In the process, they end up discovering a series of squares and, he tells us, first and second order discrete differential equation – they’re six year olds…
Showing us the Squeak computing environment, he shows us children building cars and learning about acceleration by programming them to move across the screen. In footage familiar to anyone who’s seen Alan talk, we see 11 year olds learn about acceleration by dropping balls off a school roof, filming the drops and discovering acceleration via gravity.
Kay tells us that children need mentors, like the remarkable teachers who help students learn science with these tools. The reason for creating the One Laptop project, he tells us, is that we can create five million laptops this next year, but “we couldn’t create a thousand teachers to save our lives”. Real mentoring through computers will take a reinvention of the computer interface, something he believes will cost about $100 million. It sounds like a lot, he tells us, but it’s nothing compared to our daily spending in Iraq.
I have lots of questions for Kay after this talk – is Sugar not sufficient, in his opinion, for OLPC to reach its goals? How will OLPC help create mentors? Would his re-envisioned interface mean the machine could be a mentor? This is one of those moments where I wish TED had more space for questions…