Видео

Good conversation, but I think you didn’t talk enough about scale. Waymo need to premap a city. And that can’t be automated, cause the system can’t drive there. And waymo miles and tesla fsd miles aren’t even comparable. In a chart you almost can’t see the waymo miles

Right off the bat with the SAE levels already with the bad takes lol

The competition with China started with cooperate & military espionage. Economic warfare through currency manipulation & banning many U.S. sectors from competing in their markets. The grey zone warfare in the South China Sea

we basically took a toddler, stunted its brain growth, and are now trying to proliferate clones of the toddler at every opportunity

Big pharma will swoop in snatch up all that research put a patent on it jack up the price and make it unaffordable to the average person don't get your hopes up

great work on this. your focus on solving REAL applied world practical problems sets you apart- thanks !

We need to be leaders in technology regardless of what it it. Ai and crypto are the two I think are most important right now. There will be an arms race for Bitcoin in the near future. Hopefully, Trump is able to implement his plan to make it a national strategic asset. The more crypto mining we have in the states, the better. Mining companies can be used for Ai as well. They go hand in hand👍

Using Capitalism to acheieve Marxism? Someone break down how thats possible in layman terms for me.

You have ZERO understanding of China. Stop talking.

Great content, thanks both.

ai owned by private companies? is this like neo-Jo wish communism or something? <-- see why this makes sense... ask AI...

Another great mind blackpilled by heat death. Hope he hacks his matrix soon. 💜

I mean this is an insanely awesome video. Samo is "awetastically" clear minded... his monologue about US vs. China is truly legendary... Awetastic should be a word. I just made it up.

After doing a quick research on Samo’s background-reading the pitch on Bismarck Analysis website-I was quite biased against him going into the interview. But actually listening to him argue here in person,I found myself aligned with his reasoning almost absolutely. 💁‍♂️ Good and insightful interview! Why was I biased against him? “Great Founders Theory” sounds like an oligarchic boot licking pitch deck, for which I have less than zero sympathy.

Really good talk, thank you.

That assertion re: Nordstream needs support. . .

🎉 My answer to the question of how to distinguish friend from foe is to ask, do you agree with the principles of the Universal Declaration of Human Rights and the Earth Charter? I think those principles are among the highest expressions of human intelligence to date.

Perhaps the highest priority for humanity now is building a global digital platform for collective intelligence. The platform will keep humans in the loop through a conversational interface that merges public parts of conversations into a shared graph representation. It will report on which issues are discussed by the most people, aggregate measures for sentiment about proposed solutions, and show how those measures of sentiment change over time. One big challenge is how the platform will authenticate the identities of humans at the same time that it allows people to share selected parts of conversations anonymously.

Honestly, midwit guest. Your first question was his AGI timelines to contextualize his views - he only answered it 1+ hours later. You asked him for a plausible US-China scenario besides Leopold’s - instead, he just nitpicked Leopold’s scenario in a vacuum without offering an alternative. Then he actually said, “Don’t build skynet, bro. Just don’t.” - uh, ever heard of breakout potential destabilizing arms control equilibriums? This was a frustrating listen.

I think the biggest problem with governments and A.I is the culture. There is a knee jerk reaction to punching down, and believing they understand a situation before they actually understand it. Bias in the A.I system will be taken at face value by these individuals. Also, the biggest problem, when does A.I return value to society, rather than looting the societies personal data? There still is nobody using A.I to find sustainable tech solutions, or sustainable energy solutions. As you can see, nothing good is coming from it, and states are using it to cycle wars through endless cyber terrorism. Plus when US & China cycle endless fighting, it incentives covert and adversarial surveillance networks, for powers to generate political enemies to facilitate their talking points.

damn.. he is really good at analysis 😂

Awesome video

Nice find . I'm early

These conversations need to be bigger and multidisciplinary. A version of this happened shortly after the Big Bang, during an era called the Habitable Epoch, and the energies and information from that period extend forward into the energies and information of our minds now.

1:36:14 tesla has a mobile service team that already services their entire fleet. Also charging, cleaning, etc. Tesla by far has the infrastructure advantage

Interesting definition of complexity! As mentioned, in the paper complexity is measured as the difference between Kolmogorov complexity and Shannon entropy. And measuring the Shannon entropy using bytes instead of bits I think is useful. When calculating the Shannon entropy for bits, most strings will have roughly equal number of ones and zeros resulting in very little variation between the strings whereas with bytes there will be more variation.

Very few philosophers are worth listening to at all. Unfortunately that includes this one too.

Why is he talking about afterlives? Why are there afterlives on a channel I am subscribed to? Not cool.

Aloha!

Great episode. Any chance you could share the three papers mentioned herein? I was only able to locate the "Approaching Human..." by Halawi, et al. Thanks!

Inspiring talk, thank you. I am pumped to try my hand at the tournaments. Can I use the template that is listed over at Google Collab? Can I incorporate external automation tools such as Lindy and Taskade? I have been getting nearly 2,000 t/s with Cerebras recently. Is that also available on Metaculus?

this is sooooooo good. been working on agentic workflows and this really sparks the ideas

How good is this? Seems convenient…but is it as powerful as other studios ?

Great presentation, Nathan. I love how you laid out such a clear process for approaching an automation project and included very specific, concrete advice like the 10/100/1000 examples.

23:08 This guest seems to have some fundamental misconceptions around how Tesla accrues training data. He seems to think that FSD needs to be engaged in order for it to learn. The opposite is true. Their system is trained while humans are driving. When a human takes over from FSD, this is treated as a strong signal to pay attention to how a human driver behaves. You tried to correct his misconception but he incorrectly doubled down on it.

Fantastic episode Nathan! The presentation is an absolute treasure! Thanks so much for sharing!

28:12 "National divorce" over sb1047? Really?

The interviewer should focus on interviewing instead of talking about himself.

This was fantastic. Thank you

some great practical advice, cheers

AI automation is evolving fast! Consider alternative tools that can help you optimize your AI setups with GPT-4o fine-tuning. #AIAutomation #GPT4o

Really good content

Excellent context and direction here bud 🎉 Thank you for putting this together for everyone

The intro is pretty robotic and useless. Better jump straight in, or make it simpler

What an absolutely fascinating and exciting subject. I hope the field gets all the money thrown at it!

The thing with LiDAR vs only camera is not just cost (which is absolutely massive when scaled to the millions annually) but it’s also processing speeds. If you solely focus on everything hinging on vision than everything you do goi g forward will revolve around perfecting vision. If you rely on lidar for safety than you’re handicapping yourself in terms of vision while simultaneously bogging down the entire processing network as the two systems compare every frame with one another to concur that they’re right. LiDAR is not bad by any means, but in terms of future proofing autonomy it is like a training wheel that demands more energy, cost and processing power which could otherwise be used more efficiently on purely vision. We’ll see in the long run but I imagine that LiDAR will not be used in autonomous vehicles in 5+ years as all switch to the vision based approach.

if you've got 100 million in a model and you can't figure out what's going awry, who you gonna call.. Goodfire! 😂 that's interesting though, I wasn't aware of sparse autoencoders

YOU HAVE THE BEST PODCAST ON RUclips. Edge of my seat over here with every episode. Thanks for making these!

Well clearly suffering is not a factor that institutions in America care about, they constantly harvest personal data and wage psyops and surveillance hate campaigns, I think that is the goal. So lets go full USA mode and create more hell simulations, and put every poor person on a surveillance and psyops warfare data base, and destroy everyone's ability to live a good life, and cycle wars in the middle east, and form a cold war with China.

👏