On: Some excerpts from recent Alan Kay emails

[Jyaif]

Is the underlying assumption behind this text is that less innovation is happening right now? Are the billions poured in AI not counting?

[hashmal → Jyaif]

The assumption is (I believe) that innovation is better when it's not goal-oriented. Currently much of the funding goes to projects that have a clear path to results, but finding that path is a prerequisite, and is not funded.

[Glench → Jyaif]

Kay makes a distinction between invention and innovation. Invention is the creation of qualitatively new things and innovation is the betterment of pre-existing things. The AI research that’s happening now sounds more like innovation. Innovation is fine, but civilization-changing things (like the constitution or Internet) are the result of invention, not incremental improvement (to paraphrase Kay’s argument).

[paulsutter → Jyaif]

Particularly considering that most of the AI work is openly published (companies averse to publishing have difficulty hiring in AI)

[coldtea → Jyaif]

>Are the billions poured in AI not counting?

Unless there are any great results, no, they are not.

Tons of research money where poured into AI in the 60s-70s too, and when it got nowhere great, it all folded in the so-called "AI winter".

[bboreham → Jyaif]

The comparison is with an environment where the mouse, windowing and Ethernet were created.

Billions poured into AI have produced some interesting tricks but little that changed the entire world of computing in the way those things have.

[Jyaif → bboreham]

That's funny you should mention "the mouse, windowing and Ethernet", because in the last 5 years the majority of people have stop using all 3 of those (thanks to mobile). So that takes care of the "little that changed the entire world of computing" argument.

Yes, billions have been poured into AI and we still have no strong AI. That just shows that it's a problem that is way more complicated that inventing the concept of a pointing device, making computers talk to each other, or having applications show something on a screen that isn't text. The people in the 80s weren't smarter (or dumber) than the people of today, there were just a lot of low hanging fruits back then.

[icc97 → Jyaif]

> because in the last 5 years the majority of people have stop using all 3 of those (thanks to mobile).

Even if this were true, which I disagree with, it has changed the world. Just because we don't still use the iPod anymore doesn't make it any less brilliant.

> The people in the 80s weren't smarter (or dumber) than the people of today, there were just a lot of low hanging fruits back then.

They didn't just take the low hanging fruit. In the 60s it was assumed that computers would just get bigger and bigger. They had to take the highest hanging fruit that everyone thought was ridiculous. They had to envision an entirely different world. So perhaps they weren't more intelligent but they had a more powerful vision of the future.

You're probably laughing at similar people today.

[cma → bboreham]

Things wouldn't be too different without the mouse. We already had the trackball.

[icc97 → Jyaif]

As far as I can tell, it's a bit like the 20% Google time. Except for instead of it being one day a week, it's 25 people in a building for 5 years.

The people being given money for AI are expected to work on AI, not on whatever they think is cool.