Luke Pettit

Bio…

[Luke Pettit @RealLukePettit]

@spuicey Is the checkpoints included on the free version?

[Luke Pettit @RealLukePettit]

@spuicey So the credits go towards cloud builds, sound, and other tools?

And our ChatGPT subscription can be used for the AI part?

[Luke Pettit @RealLukePettit]

@spuicey On your free version?

[Luke Pettit @RealLukePettit]

@spuicey Look how far AI has come

[Luke Pettit @RealLukePettit]

@spuicey @rileybrown @Stefan_3D_AI

[Luke Pettit @RealLukePettit]

Beautiful

[spuice @spuicey]

I can't believe how many hours I spent hand animating reloads.

This blender fork rigged and animated it from a 6 second video.

[Luke Pettit @RealLukePettit]

Real

[David Ondrej @DavidOndrej1]

i dont even check my Vercel anymore

a year ago I would babysit every deployment

now, the models just don't make mistakes

[Luke Pettit @RealLukePettit]

Once BCI become advanced…

- We use it to replicate a complete model of the brain’s neural network.

- Then predict what a person will do next based on the sensory inputs that are coming through.

Is life predictable?

[Luke Pettit @RealLukePettit]

@arlogilbert @NousResearch All they are doing is grabbing the first 11k characters and last 4k characters of the webpage and gives that to AI instead of handing off an AI summary.

Then if AI needs more it just searches through the middle.

[Luke Pettit @RealLukePettit]

@fin465 Send

[Luke Pettit @RealLukePettit]

More like 1/10 would even know the company Anduril.

Sometimes u have to step out of SF.

[Ken Wattana @KenWattana]

I think if you asked 100 people who Anduril's CEO was, 90 of them would say Palmer or Trae

[Luke Pettit @RealLukePettit]

This generation is ROTTING.

Assuming you, me, and every other person on earth scrolled shorts everyday.

That would be 24 SCROLLS/DAY…

[YouTube Creator Insider @YouTubeInsider]

Now averaging over 200 billion Shorts daily views!

[Luke Pettit @RealLukePettit]

I disagree.

LLMs can come up with new science.

There are 3 issues though:
1. Tools
2. Context
3. Long training loop

Tools - The issue is most LLMs are not equipped with necessary tools to carry out the scientific method specifically testing the hypothesis.

Context - LLMs ability to “clearly think” decreases significantly nearing context window

Long training loop - AI can’t learn in depth about any ideas or test it has done as it isn’t getting its weights adjusted. The only way is by storing this information and injecting it into the context which leads to “noisier thinking”

[Rohan Paul @rohanpaul_ai]

Columbia CS Prof explains why LLMs can’t generate new scientific ideas.

Bcz LLMs learn a structured “map”, Bayesian manifold, of known data and work well within it, but fail outside it.
But true discovery means creating new maps, which LLMs cannot do.

[Luke Pettit @RealLukePettit]

If you believe in God, you already believe we’re in a simulation.

God created (simulated) a reality governed by an external intelligence.

Whether you call it “God” or a “Programmer” is just perspective.

[Peter H. Diamandis, MD @PeterDiamandis]

YES, obviously, we are living in a simulation, but what will you differently now that you know that???

[Luke Pettit @RealLukePettit]

To re-iterate...

SpaceX is the modern day Virginia Company.

[Luke Pettit @RealLukePettit]

@thePaoloTrulli I would love to try it out

[Luke Pettit @RealLukePettit]

@shl Can I try

[Luke Pettit @RealLukePettit]

@pham_blnh Very cool! Can I see the list?

[Luke Pettit @RealLukePettit]

It’s weird to think that if we don’t…

it means we are no different than artificial intelligence

[Luke Pettit @RealLukePettit]

If we can’t predict it…

THEN WE HAVE FREE WILL

If we can predict it…

THEN WE DON’T HAVE FREE WILL