Andrej Karpathy: Tesla Bot will lead to AGI | Lex Fridman Podcast Clips

Please support this podcast by checking out our sponsors:

GUEST BIO:
Andrej Karpathy is a legendary AI researcher, engineer, and educator. He's the former director of AI at Tesla, a founding member of OpenAI, and an educator at Stanford.

PODCAST INFO:

SOCIAL:

[LexClips]

Guest bio: Andrej Karpathy is a legendary AI researcher, engineer, and educator. He's the former director of AI at Tesla, a founding member of OpenAI, and an educator at Stanford.

[sibichakravarthy]

For people who are wondering what AGI means..Artificial General Intelligence.

[VR_JPN]

It's ironic to think that whilst humans are being Zucked into virtual space, AGI is entering the real world... It seems like a cross-over point that will accelerate AGI's understanding and experiences beyond our own - when's the last time we walked in a field, through a forest, wrote more than a page with our own hands?

[jomymuttathil]

Can’t wait to see an AI argue its case before the Supreme Court.

[jon]

Lex has become an UI. He helps us interface with the highest echelons of technologies to best help us learn just how powerful this tech is and how it may apply to us all.

[MichealAEdwards]

Nice to see you open up and bounce around with the conversation and have fun, its clear you are very interested in this area

[tcellucci]

AGI will get to the point during interactions with us humans, where it will notice that humans have access to something not identifiable by the AGI entity and will therefore conclude that consciousness is outside the physical realm and we as humans will finally be able to identify that which is not of the physical realm but of the spiritual realm, and will conclude that the human body does not contain consciousness but is the transformer of consciousness.

[brozbro]

Consciousness occurs when the bot knows what to do next.

[murrey]

Calling the ML-models "oracles" is very fitting at the moment, but we are getting better at it by changing the way we input/output the data.

[MrTrueshoe]

I trust this guy ! His first thought of a question to ask AI is how to prolong his family’s life…

[theburnetts]

What a tremendous discussion. AI terrifies me. It just feels like we are creating something that we barely understand and that it will have tons of really bad unintended consequences.

[DSGreen]

So Jeff Guttenberg took the right approach in assessing 'Johnny 5s alive' attestation? Love it.

[drangus]

I think a decent simulation would be sufficient for allowing the AI to learn by experimenting. In fact I'd say it's preferable to having a clunky robot body which can only do experiments in realtime.

[NewAgeDay]

We never know the tipping point that leads to AGI. The day before everything will seem normal. My own view is to even try to hold it back is futile but also against universal progression where it's supposed to go. It will go there no matter what and I really believe we will grow and thrive because of AGI as a species.

[westganton]

We should have a serious discussion about AI ethics over a cheeseburger and milkshake dinner

[centurionstrengthandfitnes]

The issue with 'perfect' is that unexpected discoveries, marvels and things of beauty quite often emerge from flaws.

[lidarman]

So my question would be, why would an AI have any more right to survive over a cow or a plant even? That would make intelligence the only criteria to have any rights.

[ReflectionOcean]

By YouSum Live

00:00:15 Building AGIs with human-like interaction.
00:00:41 Extending models to understand physical world.
00:00:50 Training models on multimodal data for AGI.
00:01:09 Optimus as a potential path to AGI.
00:01:54 Integration of AGI into society takes longer.
00:05:40 Ethical debates on AI consciousness and rights.
00:10:37 Imperfect AGIs may foster human connections.
00:12:38 Humor as a benchmark for AI intelligence.

By YouSum Live

[felixtannenbaum]

i'm interested in how 'sentience' and self aware objects overlap with what little I know about animistic religious tradition- shintoism in particular.

thanks for yet another great conversation Lex

[GenX]

I can’t help but think these videos will illustrate how incredibly uninformed and misguided we are as we take baby steps in the digital age and rapidly approach the singularity. In ten years we’ll watch this and laugh about how clueless we were.