Artificial Intelligence

[Buckminster]

Lex had an interesting point in this video about how Neuralink could leverage the human brain for computation as well as the other way around (19 mins in I think).

The other thought I had; could a neural net learn to use a quantum computer more effectively. ie. a quantum computer as a component within a neural net.

 

[TheTalkingMule]

Is there a reason you need to bump/start all your pointless threads every weekend until the whole front page is your nonsense?

Are we going to moderate this forum or is it just dead?

 

[Buckminster]

https://twitter.com/elonmusk/status/1281661156294193152
Cool thread. Neuralink update 28th August. Will be able to cure addiction.

 

[Buckminster]

A lil birdy (called Lex) told me Jim Keller was at the Neuralink demo in person. Jim left Intel in June.

 

[SPadival]

Also, Lex said he'll have Jim on his podcast soon.

 

[Buckminster]

MODS: This may be of general interest to genpop , but off course, feel free to kick it to a sub-forum.

Thanks DD ! - great summary: Good level of detail combined with clear explanations.
Follow up question:
In similar projects of this magnitude what is the ballpark ratio of work used for data transfer, compression or bottleneck workarounds, as you describe it, compared to solving the actual problem (FSD)?
Clarification:
Did Keller really say 10- 1 million times faster/better or are you using your own background to evaluate the upper bound? Did he compare to the (in car) FSD version 3 chip efficiency, or to other ways of doing FSD (competitors)

Wild-eyed speculation: Perhaps Tenstorrent will built some kind of specialized compute unit which Tesla can use in DOJO? Perhaps made-to-order, Tesla-only. Perhaps not a main or bulk part but supplementary? I love Teslas vertical integration, but sometimes having a trusted partner is worth a lot.

General amazement:
I still find it hard to believe that DOJO is generic enough for other tasks. Elon said it could mine bitcoin, and your summary seems to imply a generic quality. To my limited understanding, FSD is (or was) considered so freaking hard that solving that requires specialized hardware - as evidenced by Tesla doing exactly that re. their custom car chips.
It also kind of doesn't make sense financially: Solving FSD is worth so much money, that even making a lot of bitcoins wouldn't really measure up. On the other hand, if solving FSD takes a number of years, and a huge amount of compute and custom chips, having some extra income is useful.

The only way that I can make sense of DOJO being generic is if DOJO is actually a trojan horse kind of tech for solving AGI !!
Is there a chance that this is actually what Tesla is trying to do? Or am I flying of a tangent here?
(How does that rhyme with Elons continued warnings about AI?)

Or, if DOJO does not solve AGI entirely, then solves an at least a large subset of AGI. Or perhaps doesn't quite solve, but boosts other known techniques by a significant order of magnitude - a kind of AGI runway. Which in the end may be the missing link for solving AGI ?..!
Maybe Elon concluded that FSD was close enough to AGI that he might as well solve for AGI. And then get FSD 'for free'.
If that is the case, then solving for a subset of AGI is worth a ... what is the level above f***ton?
Most of us are getting used to Tesla being 10+ startups. I used to think that solving FSD was worth a lot - to mankind, but also to us fans and investors. Solving AGI (or a significant subset) Danm! That is is 'huger than huge'.

If true, then someone PM Warren Redlich - his most crazy estimates are way too conservative!

(AGI: Artificial General Intelligence)

 

[Buckminster]

Karpathy podcast:
The Robot Brains Podcast

 

[heltok]

Karpathy podcast:
The Robot Brains Podcast

Awesome, thanks!

 

[Buckminster]

Are we witnessing the dawn of post-theory science?

 

[Buckminster]

Liked by Elon:

https://twitter.com/x/status/1617082270501965826

 

[Buckminster]

Tesla's former head of AI joins Elon Musk-founded OpenAI
I predict Karpathy will return to Tesla in 2025. Bard has shown that Openai know what they are doing so he might return with some useful knowledge.

 

[Buckminster]

We don't have GPT4 yet:

https://twitter.com/x/status/1627189548395503616

 

[Buckminster]

https://twitter.com/x/status/1631041849052782592

API cost down.

 

[Buckminster]

https://twitter.com/x/status/1635348855473721346

 

[Buckminster]

GPT4 Youtube link

 

[heltok]

I recommend reading this old post:

The state of Computer Vision and AI: we are really, really far away.

Musings of a Computer Scientist.

karpathy.github.io

And then reflect on this:

And think about what this says about where we will be in 11 more years...

 

[Buckminster]

That was fast:

https://twitter.com/x/status/1636377233018490885

 

[Baumisch]

puh - something is brewing quick ... I can't get this picture out of my head:

SRC: The Artificial Intelligence Revolution: Part 1 - Wait But Why

 

[Baumisch]

I recommend reading this old post:

The state of Computer Vision and AI: we are really, really far away.

Musings of a Computer Scientist.

karpathy.github.io

And then reflect on this:
View attachment 917970

And think about what this says about where we will be in 11 more years...

tbh - OpenAI dissabled external references (at least images) a few days ago - so we can't test on Andre's image nor can we check if the reddit post is real I hope they re-enable image urls again in the future. GPT 3.5 in Playground can parse image URLs still but the results are ... not good

 

[heltok]

tbh - OpenAI dissabled external references (at least images) a few days ago - so we can't test on Andre's image nor can we check if the reddit post is real I hope they re-enable image urls again in the future. GPT 3.5 in Playground can parse image URLs still but the results are ... not good

The image is from this paper:

https://cdn.openai.com/papers/gpt-4.pdf

(page 9)

There are many more examples later in the paper, all equally impressive.