Yes, there was a certain golden period of gp's building on Koza and others
work like Forrest Bennett's Beowulf 1000-pentium cluster back in the late
90s.
https://www.genetic-programming.com/machine1000.html
I agree with you, Marcus, that it would be good to see versions of this
springing forth
I don’t understand why Genetic Programming hasn’t been a bigger thing. It
seems like another case, like ML, where having adequate hardware is key to
really making it work. I hope interest in AI will dust-off or reinvent many
such approaches.I don’t care who gets the credit.
From: Friam
I see I missed this on HackerNews yesterday, the comments cover a range of
positions.
https://news.ycombinator.com/item?id=38642651
-- rec --
On Fri, Dec 15, 2023 at 12:48 PM Roger Critchlow wrote:
> On the subject of cognitive dissonance, and working for large research
> institutions
>
>
As I recall, CMU had a strong AI group before Hinton got there. Raj Reddy,
Scott Fahlman, Marc Raibert, etc. I wondered why the media called him "The
Godfather".
On Fri, Dec 15, 2023 at 12:48 PM Roger Critchlow wrote:
> On the subject of cognitive dissonance, and working for large research
>
On the subject of cognitive dissonance, and working for large research
institutions
https://people.idsia.ch/~juergen/ai-priority-disputes.html
in which Jürgen Schmidhuber complains that LeCun, Bengio, and Hinton claim
to have invented modern AI when they were actually mostly reinventing,
While I think LLMs will be hard to use for analysis tasks, there is something
satisfying to see certain people squirm as LLMs, as Altman says, “Blow right
through the Turing test.”
From: Friam On Behalf Of Roger Frye
Sent: Wednesday, December 13, 2023 6:47 AM
To: The Friday Morning Applied
