LLMs will be democratized when Apple releases the Apple Silicon M99 chip with 100,000 cores in a machine for about $549. At the moment, M2 provides only a measly 16 neural cores.

Perhaps a Galilean inversion of Ptolemaic Software Cosmology will be required?

Galilean Software Cosmology: The Universe does not revolve around a solitary CPU chip, but, is composed of zillions of CPU chips.

In the 1950s, it was believed that The CPU was the center of the Universe - i.e. Ptolemaic Software Cosmology.

In 2023, the ground rules have fundamentally changed. Programmers have bowls full of $1 CPUs and essentially-infinite amounts of memory.

For example, VISA gave me - for free - a plastic card with an embedded CPU plus local persistent memory. It doesn’t even need batteries.

For a further example, every person carries around super-computers in their pockets these days, each being connected to zillions of other super-computers via cellular networks.

The problem is no longer “how to make a single CPU run faster”, but, has become “how to make a network of zillions of nodes more manageable and to run faster and to be cheaper”. Ptolemaic Software Cosmology is unsuited to dealing with such (new) problems.

See Also

References

https://guitarvydas.github.io/2021/12/15/References.html

Blogs

blog

obsidian blogs (see blogs that begin with a date 202x-xx-xx-)

Videos

videos - programming simplicity playlist

Books

leanpub’ed (disclaimer: leanpub encourages publishing books before they are finalized - these books are WIPs)
Programming Simplicity Takeaways, and, Programming Simplicity Broad Brush

Discord

Programming Simplicity all welcome, I invite more discussion of these topics, esp. regarding Drawware and 0D

Twitter

@paul_tarvydas

Mastodon

(tbd, advice needed re. most appropriate server(s))