Google's New AI Feature is UNREAL...

The latest AI News. Learn about LLMs, Gen AI and get ready for the rollout of AGI. Wes Roth covers the latest happenings in the world of OpenAI, Google, Anthropic, NVIDIA and Open Source AI.

My Links 🔗

#ai #openai #llm

LINKS:

Notebook LM

GameGen-O

How ARK Is Thinking About Humanoid Robotics

[torq]

I was surprised how little time it took for me to forget I was listening to an AI conversation.

[fiery_transition]

the podcast itself without the video, breaking down topics into a more natural flow of investigative questioning and answering? This is huge for learning or digesting stuff!

[AzzaTwirre]

The backwards walking was a clever touch

[jhdesigner]

This is the worst this Tech will ever be.... In 2 years looking back... this will look awful!

[EnslaveTV]

The irony of an ai podcast host on an ai "show" on youtube discussing if robots will take their jobs. Thats so meta

[yazuki-wolf]

Damn, I meant to click off the video after a few minutes of the ending podcast, just to get the gist of it, but I ended up listening to the whole thing and felt like I wanted more by the end. This is dangerously good.

[HappyCodingZX]

This gives me the same feeling as when a random person approaches you in the street in a friendly way and underneath you know something isn't quite right about them. AI has evolved to the point where it can accurately emulate creepy people.

[iham]

next step: make video podcasts on school topics; give students the ability to choose their favorite characters and voices per topic; add scenery if useful (history of rome - show rome); make it interactive by letting students ask questions. school could go in a direction everybody enjoys it as if it is a game/documentary and create a positive emotion and bond to all sorts of topics.

[RickeyBowers]

The last industrial robot I worked with was 3/4 million USD. That's because there is a huge capital investment to create the environment for the robot. In the end - the robot was slower than a human, but could work 24/7 with near perfect consistency - the output of two humans. Over five years this resulted in a net profit.

[BerchtwaldEliasEH]

Every crash/collapse brings with it an equivalent market chance if you are early informed and equipped, I've seen folks amass up to $1m amid economy crisis, and even pull it off easily in favorable conditions. Unequivocally, the collapse is getting somebody somewhere rich.

[diga]

THIS ENTIRE EXPERIENCE IS SURREAL - we live in a simulation.

[drunktrump]

well THAT was a quick pass through the Uncanny Valley. And... here we are on the other side.

[doingtime]

Feed text and media into a container and 2 minutes later get a podcast about it? I'm sorry but this is mindblowing.

We are so prone to trivialize everything but tell about this tech to someone just 5 years ago and they would say you are crazy.

[slls]

It’s funny you brought up UBI halfway through. I’ve been testing o1 with a pretty technical UBI related project.

So while watching your video I grabbed my latest draft, went to Notebook and generated almost 10 minutes of audio in one click with zero prompting or input other than the draft document itself.

The results were pretty impressive. Only two minor and relatively logical/accurate hallucinations, and somewhat misfocused in areas. But pretty incredible first attempt, especially since like I said, no direction or input was given - I just pasted 20 pages of text into notebook (losing most of the formatting), and hit the load button, came back and watched the video for another minute, went back to Notebook and there it was.

Learn to garden folks, ‘cuz not many of us will have jobs in 10 years 😂

[DeusAxios]

At least every week there is a massive breakthrough in AI. The speed of advancement is crazy.

[wylhias]

I've noticed a trend watching all those Ai news and videos. Every couple of weeks I find myself thinking "WTF"with more and more frequency with how baffling all of this is

[yesthatkarim]

So what studies have found is that it’s possible for large language models to listen to YouTube videos, and then add a comment about the video. You could potentially have an AI _commenting_ on a video generated by an AI.

Once the AI commenter has been trained on a meme dataset, its responses would never gonna give you up, never gonna let you down, never gonna r *SYNTAX ERROR IN LINE 10*

[ImpendingDoon]

Thanks for the informative video! I've used Notebook before, but they didn’t have the Deep Dive feature. Super cool. Question: How did you create the video for the podcast? I went to HeyGen and didn't see how to create a two-person video. Or did you make that video yourself? Pulling in both avatars and working with the timeline to align the character switches to the voices?

[alex.arango]

I’m not against AI, I’m against making AI look like humans. It doesn’t need to look or act like us, it needs to help us not replace us

[madaxe]

I consider myself very tech savvy. Grew up programming basic on my C64 and has kept up ever since.

This is like 1-2 steps away from me not being able to tell if it's real. Put the characters in a podcast studio, tweak the dialog a bit and pick avatars and voices that go together, and I'm not sure I could tell