OpenAI’s New ChatGPT Goes On Steroids!

That sounds awesome. I really like the idea of a scenic drive. What route should I take for a scenic drive? Taking Highway 1 is a fantastic choice. You could start in San Francisco.

Yes, OpenAI's new ChatGPT advanced voice mode is here and it is a great advisor or partner, but it is so much more than that. It opens up a whole new world of possibilities. How?

What does Claude have that I don't? Tell me what's wrong with me?

Well, for instance, you can give it a face and even gestures by combining it with two other tools. Wow. So understandably, I was so excited to try it until, uh oh, not in the eu. Well, there's that. And the good news are not over. Here is a new thing that talks that is going to put me out of business. And third, this is Google DeepMind's idea of the road to super intelligence. It might come sooner than we think, so make it two things that put me out of business. What a time to be alive. And you can think of the new ChatGPT not just as something, but as someone put it next to you when you're working and start brainstorming together. And I am really excited for the future version that will not just hear you and answer, but see too. So when you study with it, it gives you feedback in real time. Becoming a private tutor. Earlier they showcased how it can even prep you for an interview. Can you tell me about a challenging.

Technical problem you solved in your previous projects and how you approached it?

Or it can talk to you sarcastically.

I'm so excited for this.

Now we're talking real value here. Love it. You can also use ChatGPT's new voice mode for real time translation when you are in a foreign country. And interestingly, you can interrupt it.

She only speaks Mandarin. Got it.

But it might interrupt you too. Ha. You'll love it. But currently it can also sing a duet with you where you sing a few lines and then it sings a few lines and it enters at the right moment. I am showing to you all the sounds. I am allowed for this song because of YouTube's policies. Yes, nothing but this one I can show you. I love how they put two of them together to sing a song line by line.

Bunny ears and laughter, a swift retreat.

A moment of joy and the stylish sweet.

And it is just as awkward as you would expect. Two humans doing it. And now many of you can try it yourself too. This is an incredible leap forward. It's not doing my laundry yet, but I'm excited to try it someday because it has everything. Low latency, it talks fluidly. And check this out. If you don't like how it talks, just ask it to speak differently. Although this British voice is so good, you can basically forget about the rest.

It's lovely to meet you. If there's anything you need, please don't hesitate to let me know.

It could narrate my entire life. This is where he loses his papers again. And now hold on to your papers, fellow scholars, for the thing that is going to put me out of business. This is a new technique that takes a deeply technical blog post and creates a podcast out of it with two hosts and a little banter too.

We're not really dealing with accounts like in traditional finance. Instead we've got these things called UTXOs.

UTXOs sounds a little intimidating, I gotta admit.

Hmm. Sounds really natural. So good. But it gets better. You can give it something really dry. Oh, yes, this piece of computer code will do. And you get an entertaining podcast that describes it too.

Okay, so remember all those numbers we were talking about?

Yeah.

Those actually represent probabilities.

That is just mind blowing. And of course, it works for technical papers and books too. I added one of my papers on ray tracing that is in a super specialized sub area with very few scientists. And while we are waiting for the podcast edition, the summary is fantastic. FAQ also fantastic. And now. Oh yes, hold on to your papers, fellow scholars, because the podcast is done. Let them talk about it for a bit and check the contrast between how advanced the paper itself is and listen to how well they are explaining it.

So it's evaluating each step.

Exactly.

So like if a firefly lands on a super bright surface.

Right.

It's kind of like, ooh, jackpot. This is important information. Exactly. It'll probably hang out there longer.

More likely to stay there.

Yeah, but if it's in a dark corner, it's like, this is a waste of time.

Exactly.

Let's bounce and try a new spot.

That is a fantastic way to put it.

I've never seen anything like this in my life. My goodness. Wow. It is completely going to revolutionize learning and put me out of business. Thanks. AI, a true powerhouse for you, fellow scholars. It is called NotebookLM. The link is available below and an open source variant is already available. Insanity. Now on to superintelligence. Get this, Google DeepMind started to use AI to design their new chips, which are of course also running AI. And yes, I'll keep depicting every robot as a strawberry. Get used to it, fellow scholars. That is already super interesting. But it gets better. Just think about it. Their CEO Demis Hassabis says that AI designs better chips. Those chips are then used to train better AIs which will then be able to design better chips and the cycle goes on in a feedback loop. That is an absolutely amazing concept and I would love to see how it comes to fruition. So everything is amazing and wonderful and now I hope you excuse me as I gotta go get hired to Nvidia to program strawberry shaped robots or I don't know, I might need to use this for an interview prep right now. Until then, yes, check out our sponsor to support us. If you feel inspired and you wish to run your own experiments on Nvidia GPUs, check out Lambda Lambda is a cloud service for large scale AI projects and it is my absolute favorite. Seriously. Try it out now@lambdalabs.com papers or click the link in the description. If you use Lambda you can run your own experiments on super powerful hardware and you also support two minute papers.